Molecular Phylogenetics and Evolution 65 (2012) 992–1003 Contents lists available at SciVerse ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Phylogeny of bent-toed geckos (Cyrtodactylus) reveals a west to east pattern of diversification Perry L. Wood Jr. 1, Matthew P. Heinicke ⇑, Todd R. Jackman, Aaron M. Bauer Department of Biology, 147 Mendel Hall, Villanova University, 800 Lancaster Ave., Villanova, PA 19085, USA a r t i c l e i n f o Article history: Received 7 May 2012 Revised 24 August 2012 Accepted 27 August 2012 Available online 13 September 2012 Keywords: Sundaland Indonesia Biogeography Dispersal Myanmar Papuan region a b s t r a c t The Asian/Pacific genus Cyrtodactylus is the most diverse and among the most widely distributed genera of geckos, and more species are continually being discovered. Major patterns in the evolutionary history of Cyrtodactylus have remained largely unknown because no published study has broadly sampled across the geographic range and morphological diversity of the genus. We assembled a data set including sequences from one mitochondrial and three nuclear loci for 68 Cyrtodactylus and 20 other gekkotan species to infer phylogenetic relationships within the genus and identify major biogeographic patterns. Our results indicate that Cyrtodactylus is monophyletic, but only if the Indian/Sri Lankan species sometimes recognized as Geckoella are included. Basal divergences divide Cyrtodactylus into three well-supported groups: the single species C. tibetanus, a clade of Myanmar/southern Himalayan species, and a large clade including all other Cyrtodactylus plus Geckoella. Within the largest major clade are several well-supported subclades, with separate subclades being most diverse in Thailand, Eastern Indochina, the Sunda region, the Papuan region, and the Philippines, respectively. The phylogenetic results, along with molecular clock and ancestral area analyses, show Cyrtodactylus to have originated in the circum-Himalayan region just after the Cretaceous/Paleogene boundary, with a generally west to east pattern of colonization and diversification progressing through the Cenozoic. Wallacean species are derived from within a Sundaland radiation, the Philippines were colonized from Borneo, and Australia was colonized twice, once via New Guinea and once via the Lesser Sundas. Overall, these results are consistent with past suggestions of a Palearctic origin for Cyrtodactylus, and highlight the key role of geography in diversification of the genus. ! 2012 Elsevier Inc. All rights reserved. 1. Introduction Cyrtodactylus Gray, 1827 (bent-toed geckos), with more than 150 recognized species, is by far the most species-rich genus of gekkotan lizards (Uetz, 2012). Recently, as many as 19 new species have been described in a given year from throughout the group’s broad range in Asia and the western Pacific (Fig. 1), and since the start of the 21st century, known diversity of Cyrtodactylus has more than doubled. Virtually all regions occupied by Cyrtodactylus have seen a huge increase in the number of recognized species described, including Myanmar (Bauer, 2002, 2003; Mahony, 2009), Vietnam, (Ngo, 2008; Ngo and Bauer, 2008; Nguyen et al., 2006), Sundaland (Chan and Norhayati, 2010; Grismer et al., 2010; Iskandar et al., 2011; Oliver et al., 2009), the Philippines (Welton et al., ⇑ Corresponding author. Present address: Department of Natural Sciences, University of Michigan-Dearborn, 125 Science Building, 4901 Evergreen Road, Dearborn, Michigan 48128, USA. E-mail addresses: perryleewoodjr@gmail.com (P.L. Wood Jr.), heinicke@umd. umich.edu (M.P. Heinicke), todd.jackman@villanova.edu (T.R. Jackman), aaron. bauer@villanova.edu (A.M. Bauer). 1 Present address: Department of Biology, Brigham Young University, Provo, UT 84602, USA. 1055-7903/$ - see front matter ! 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ympev.2012.08.025 2009, 2010a, 2010b); and Wallacea, Australia, and New Guinea (Kraus, 2008; Oliver et al., 2011; Shea et al., 2012). Despite great activity in terms of alpha systematic studies, phylogenetic research on Cyrtodactylus has been relatively limited thus far. Regional phylogenies with reasonably broad taxon sampling have been generated for the bent-toed geckos of the Philippines (Siler et al., 2010; Welton et al., 2010a, 2010b), Australia and Melanesia (Shea et al., 2012), and for some Malay Peninsula and Sundaland Cyrtodactylus (Grismer et al., 2010). Monophyly and interrelationships of the bent-toed geckos of these and other geographic regions has yet to be established. Indeed, the composition of Cyrtodactylus as a whole remains unclear, especially with respect to certain taxa in Nepal, northern India, and Tibet, which have been variously assigned to Cyrtodactylus, Cyrtopodion, Altigekko, Altiphylax, Indogekko, and Siwaligekko (see Krysko et al. (2007) for a recent review). In addition, the status of Geckoella, a presumably monophyletic group of small, ground-dwelling bent-toed geckos (regarded as either a genus or a subgenus of Cyrtodactylus; Kluge, 2001; Bauer, 2002) remains uncertain. A practical problem engendered by the lack of broader scale phylogenetic resolution in Cyrtodactylus is that each newly described species must be diagnosed relative to all of its congeners, P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 993 Fig. 1. Global distribution of Cyrtodactylus, with place names mentioned in the text listed. or the assumption must be made that geographically coherent and morphologically similar species are monophyletic. While this appears to be the case in some instances, it is likely that not all such groups are natural, and without broad sampling monophyly cannot be conclusively demonstrated even for regional clades that do appear monophyletic (e.g., Philippines: Siler et al., 2010; Welton et al., 2010a, 2010b; Queensland: Shea et al., 2012). A phylogeny spanning the geographic distribution and morphological variation of Cyrtodactylus and having sufficient sampling to provide a ‘‘backbone’’ for the genus would make a major contribution in providing a preliminary estimate of the monophyly of as yet unevaluated presumptive clades. In addition, such a phylogeny can provide information about relevant outgroups for future regional Cyrtodactylus phylogenies, allow evaluation of the generic allocation of the problematic taxa that have been variously assigned to Cyrtodactylus or other genera, and provide a framework for comparative analyses of Cyrtodactylus biology, including biogeography and morphological evolution. To this end we used nucleotide sequence data from approximately 45% of recognized bent-toed gecko species, including exemplars encompassing the morphological range of variation and from across the geographic range of Cyrtodactylus, to erect such a backbone phylogeny for the genus. We use this data set to evaluate current taxonomy and make preliminary observations of Cyrtodactylus historical biogeography. 2. Materials and methods 2.1. Taxon sampling, data collection, and sequence alignment We constructed a nucleotide sequence dataset for 68 species of Cyrtodactylus from throughout the range of the genus, plus two species of Geckoella, six Hemidactylus, and six genera and species of Palearctic naked-toed gecko. For purposes of establishing a timetree for the group, we included six additional gekkotan outgroups as well as Anolis and Python (Table 1). Three outgroups – Python, Lialis, and Stenodactylus – are composites of two closely related species. The dataset consists of the complete mitochondrial gene ND2 and flanking tRNAs (Ala, Asn, Cys, Tyr), plus portions of the nuclear genes RAG1, PDC, and MXRA5. New sequences are deposited under GenBank accession numbers JX440515–JX440726. Liver, muscle, or tail tissue samples were derived from individuals collected in the field by the authors or donated by other researchers (see acknowledgments). When possible, specimens themselves or photographic vouchers were examined by one or more authors but in some cases we were dependent on the species identifications of collectors or other institutions. Given the rapid rate of description of Cyrtodactylus spp. and the break-up of ‘‘species’’ previously believed to be widespread (e.g. Johnson et al., 2012), it is possible that some identifications may need subsequent revision, however, all may be considered accurate to at least species group. Genomic DNA was extracted from tissue samples using Qiagen DNeasy™ tissue kits under manufacturers’ protocols. All genes were amplified using a double-stranded Polymerase Chain Reaction (PCR). Included in the reaction were 2.5 ll genomic DNA, 2.5 ll light strand primer 2.5 ll heavy strand primer, 2.5 ll dinucleotide pairs, 2.5 ll 5! buffer, MgCl 10! buffer, 0.18 ll Taq polymerase, and 9.82 ll H2O, using primers listed in Table 2. PCR reactions were executed on an Eppendorf Mastercycler gradient theromocycler under the following conditions: initial denaturation at 95 "C for 2 min, followed by a second denaturation at 95 "C for 35 s, annealing at 50–55 "C for 35 s, followed by a cycle extension at 72 "C for 35 s, for 34 cycles. All PCR products were visualized via 1.5% agarose gel electrophoresis. Successful PCR amplifications were purified using AMPure magnetic bead solution (Agencourt Bioscience). Purified PCR products were sequenced using Applied Biosystems BigDye™ Terminator v3.1 Cycle Sequencing ready reaction kit or DYEnamic™ ET Dye Terminator kit (GE Healthcare). Products were purified using a Cleanseq magnetic bead solution (Agentcourt Bioscience). Purified sequence reactions were analyzed using an ABI 3700 or ABI 3730XL automated sequencer. All sequences were analyzed from the 30 and the 50 ends independently to ensure congruence between the reads. The forward and the reverse sequences were imported and edited in Geneious™ version v5.4 (Drummond et al., 2011); ambiguous bases were corrected by eye. All edited sequences were aligned by eye. Protein-coding sequences were investigated in MacClade v4.08 (Maddison and Maddison, 2003) to ensure the lack of premature stop codons and to calculate the correct amino acid reading frame. 2.2. Phylogeny reconstruction For comparative purposes, phylogenetic reconstructions were implemented using one character-based approach, Maximum Parsimony (MP) and two model-based approaches, Maximum Likelihood (ML) and Bayesian Inference (BI). Maximum Parsimony (MP) phylogeny and bootstrap estimates for nodal support were ID number Species Anolis carolinensis Python molurus Python regius Agamura persica Bunopus tuberculatus Bunopus tuberculatus Cyrtopodion elongatum Hemidactylus anamallensis Hemidactylus angulatus Hemidactylus frenatus Hemidactylus frenatus Hemidactylus garnotii Hemidactylus garnotii Hemidactylus mabouia Hemidactylus turcicus Lialis burtonis Lialis burtonis Lialis jicari Mediodactylus russowii Oedura marmorata Pygopus nigriceps Sphaerodactylus roosevelti Sphaerodactylus torrei Stenodactylus petrii Stenodactylus petrii Stenodactylus slevini Tropiocolotes steudneri Woodworthia maculata Cyrtodactylus (Geckoella) deccanensis Cyrtodactylus (Geckoella) triedra Cyrtodactylus adorus Cyrtodactylus agusanensis Cyrtodactylus angularis Cyrtodactylus annandalei ACD 2637 Cyrtodactylus annulatus LSUHC 7286 CAS 216459 Cyrtodactylus aurensis Cyrtodactylus ayeyarwadyensis SP 06906 Cyrtodactylus baluensis LSUHC 8933 CAS 214104 Cyrtodactylus batucolus Cyrtodactylus brevidactylus LSUHC 4056 CUMZ 2003.62 Cyrtodactylus cavernicolus Cyrtodactylus chanhomeae n/a n/a n/a Pakistan, Balochistan, Makran district, Gwadar division United Arab Emirates, Sharjah Iran, Qeshm Island captive India, Tamil Nadu, Ervikulam Ghana, Volta region, Togo Hills Myanmar, Tanintharyi Division, Kaw Thaung District New Caledonia, Sommet Poum Myanmar, Rakhine State, Taung Gok Township New Caledonia, Sommet Poum USA, Florida USA, Louisiana Australia, Victoria, Beulah Station Australia Australia captive Australia, Queensland Australia, Northern Territory USA, Puerto Rico Cuba captive Niger, 49 km S Agadez Saudi Arabia, Ibex Reserve captive New Zealand, Titahi Bay captive (from Indian stock) Sri Lanka, Yakkunehela Australia, Northeast Queensland Philippines, Dinagat Island, Municipality of Loreto Thailand, Sa Kaeo, Muang Sa Kaeo Myanmar, Sagging Division, Alaung Daw Kathapa National Park, Gon Nyin Bin Camp Philippines, Mindanao Island, Eastern Mindanao, Diwata Mountain Range West Malaysia, Johor, Pulau Aur, behind kg. Berhala Myanmar, Rakhine State, Than Dawe District, Gwa Township, Rakhine Yoma Elephant Range, Elephant Camp Malaysia, Borneo, Sabah, Mt. Kinabalu National Park, Headquarters West Malaysia, Melaka, Pulau Besar Myanmar, Mandalay Division, Popa Mountain Park, Kyauk Pan Tawn Township East Malaysia, Sarawak, Niah Cave Thailand, Saraburi Province, Phraputthabata District, Khun Khlon Subdistrict, Thep Nimit Cave GenBank accession numbers ND2 MXRA5 PDC RAG1 EU747728 AAWZ 02008741 AEQU 010243110 AAWZ 02013979 AEQU 01027927 AAWZ 02015549 AEQU 010344888 AB177878 JX440515 JX440566 JX440625 JQ945355 JX440675 JX440676 HQ443540 JX440516 JX440567 JX440626 HM662368 EU268336 HM559662 JX440677 HM622353 EU268306 HM559695 EU268332 EU268302 HM559672 EU268329 HM559705 EU268299 GU459742 GU459540 JX440627 EF534819 EF534823 EF534825 EF534829 JX440628 JX440678 EF534779 EF534783 EF534785 EF534788 JX440679 JX440629 GU459651 JX440630 JX440631 JX440680 GU459449 JX440681 JX440682 JX440632 JX440633 JQ945301 JX440683 JX440580 JX440581 JX440634 JX440684 JX440685 JQ889178 JX440527 JX440582 JX440583 JX440635 JX440636 JX440686 JX440687 JX440528 JX440529 JX440584 JX440637 JX440688 EU268367 HM559629 JX440568 EU268363 JX440569 HM559639 EU268360 JX440570 AY369025 JX440517 GU459951 JX440518 JN393943 JX440519 JX440571 JX440572 JX440573 JX440574 HQ443548 JX440520 GU459852 JX440521 JX440522 HQ401166 GU550818 JX440523 JX440524 JX440575 JX440576 JX440577 JX440578 JX440579 GU366085 JX440525 JX440526 GU366080 P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 n/a n/a n/a FMNH 247474 CAS 228737 MVZ 234350 JB 127 CES 08022 MVZ 245438 CAS 229633 AMS R167808 CAS 223286 AMS R167800 YPM 14798 LSUMZ H-1981 AMS 141027 JFBM 8 n/a n/a AMS 143861 MVZ 197233 CAS 198428 JB 34 JB 35 MVZ 238919 MAM 3066 JB 28 RAH 292 JB 7 AdS 35 n/a KU 310100 FMNH 265815 CAS 215722 Locality 994 Table 1 Specimens used for phylogenetic analyses in this study. Identification numbers are abbreviated as follows: ACD, Arvin C. Diesmos field collection; AdS, Anslem de Silva field series (specimens pending accession at the National Museum of Sri Lanka); AMS, Australian Museum, Sydney; BPBM, Bernice P. Bishop Museum; CAS, California Academy of Sciences; CES, Center for Ecological Sciences, Indian Institute of Sciences, Bangalore; CJS, Christopher J. Schneider field series; CUMZ, Chulalongkorn University Museum of Zoology; FK, Fred Kraus field series; FMNH, Field Museum of Natural History; ID, Indraneil Das field series; IRSNB, Institute des Sciences Naturelles du Belgique, Brussels; JB, Jon Boone captive collection; JFBM, James Ford Bell Museum of Natural History (Minnesota); KU, Kansas University Museum of Natural History; LSUHC, La Sierra University Herpetological Collection; LSUMZ, Louisiana State University Museum of Zoology; MAM, Mohammed Al-Mutairi field series; MFA, M. Firoz Ahmed field series; MVZ, Museum of Vertebrate Zoology (Berkeley); RAH, Rod A. Hitchmough field series; RMBR, Rafe M. Brown field series; SP, Sabah Parks Reference Collection; TNHC, Texas Natural History Collection; USNM, United States National Museum (Smithsonian); WAM, Western Australian Museum; YPM, Yale Peabody Museum; ZRC, Zoological Reference Collection, Raffles Museum. Cyrtodactylus chrysopylos LSUHC 8595 Cyrtodactylus cf. condorensis LSUHC 6546 WAM R98393 LSUHC 8598 Cyrtodactylus consobrinus Cyrtodactylus darmandvillei Cyrtodactylus eisenmanae LSUHC 6471 BPBM 18654 CES 091196 USNM 559805 Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus CAS 222412 Cyrtodactylus gansi LSUHC 8638 LSUHC 8583 Cyrtodactylus grismeri Cyrtodactylus hontreensis n/a FMNH 255454 FMNH 265812 FMNH 258697 KU 314793 Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus FMNH 255472 MVZ 239337 Cyrtodactylus jarujini Cyrtodactylus jellesmae MFA 50083 WAM R164144 n/a FK 7709 Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus khasiensis kimberleyensis klugei loriae n/a LSUHC 7532 ID 8424 RMBR 00866 Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus louisiadensis macrotuberculatus malayanus malayanus ABTC 48075 TNHC 59549 n/a BPBM 23316 JB 126 LSUHC 8906 LSUHC 8672 CUMZ R2005.07.30.54 FMNH 236073 n/a LSUHC 4069 LSUHC 6637 LSUHC 6729 LSUHC 4813 KU 309330 BPBM 19731 CAS 226137 Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus marmoratus ‘‘marmoratus’’ mcdonaldi novaeguineae oldhami pantiensis paradoxus peguensis philippinicus pronarus pubisulcus pulchellus pulchellus quadrivirgatus redimiculus robustus russelli AMS R134930 Cyrtodactylus salomonensis elok epiroticus fasciolatus feae hoskini interdigitalis intermedius irregularis jambangan Myanmar, Shan State, Ywa Ngan Township, PanlaungPyadalin Cave Wildlife Sanctuary Vietnam, Kien Giang Province, Kien Hai District, Hon Son Island West Malaysia, Selangor, Kepong, FRIM Indonesia, Gua, 7 km NW Sumbawa Besar Vietnam, Kien Giang Province, Kien Hai District, Hon Son Island West Malaysia, Pahang, Fraser’s Hill, the Gap Papua New Guinea, Morobe Province, Apele, Mt. Shungol India, Uttarkhand, Mussoorie Myanmar,Mandalay Division, Popa (village), vicinity of Popa Mountain Park Myanmar, Chin State, Min Dat District, Min Dat Township, Che stream Vietnam, An Giang Province, Tuc Dup Hill Vietnam, Kien Giang Province, Kien Hai District, Hon Tre Island Australia, Northeast Queensland Lao PDR, Khammouan Province, Nakai District Thailand, Sa Kaeo, Muang Sa Kaeo Lao PDR, Champasak Province, Pakxong District Philippines, Mindanao Island, Zamboanga del Sur Prov., Municipality of Pasonanca, Pasonanca Natural Park, Tumaga River Lao PDR, Bolikhamxay Province, Thaphabat District Indonesia, Sulawesi Island, Propinsi Sulawesi Selatan, Kabupaten Luwu Utara, Kecematan Malili, ca. 4 km N of Malili India, Assam, Kaziranga, Kohora, Haldhibari Australia, Western Australia, East Montalivet Island Papua New Guinea, Sudest Island Papua New Guinea, Milne Bay Prov., Bunisi, N slope Mt. Simpson Papua New Guinea, Sudest Island Malaysia, Kedah, Pulau Langkawi, Gunung Raya Malaysia, Sarawak, Gunung Mulu National Park Indonesia, Borneo, Kalimantan, Bukit Baka Bukit Raya National Park Indonesia, Java Indonesia, Propinsi Maluku, Buru Island, Dusun Labuan Australia, Northeast Queensland Papua New Guinea, West Sepik Prov., Parkop, Toricelli Mts. captive West Malaysia, Johor, Gunung Panti FR, Bunker Trail Vietnam, Hon Nghe Island Thailand, Khao Luang National Park Philippines, Romblon Island Australia, Northeast Queensland East Malaysia, Sarawak, Niah Cave West Malaysia, Selangor, Genting Highlands West Malaysia, Penang, Pulau Penang, Moongate Trail West Malaysia, Pahang, Pulau Tioman, Tekek-Juara Trail Philippines, Palawan Island, Municipality of Brooke’s Point Papua New Guinea, Sudest Island Myanmar, Sagaing Division, Hkamti Township, Htamanthi Wildlife Sanctuary, upper Nat E-Su stream Solomon Islands, New Georgia I., Mt Javi, 5 km N Tatutiva Village, Marovoa JX440530 JX440585 JX440638 JX440689 JX440531 JX440586 JX440639 JX440690 JX440532 JX440533 JX440534 JX440587 JX440588 JX440589 JX440640 JX440641 JX440642 JX440691 JX440692 JX440693 JQ889180 JX440535 JX440590 JX440591 JX440536 JX440592 JX440643 JX440644 HM622366 JX440645 JX440694 JX440695 HM622351 JX440696 JX440537 JX440593 JX440646 JX440697 JX440538 JX440539 JX440594 JX440595 JX440647 JX440698 JX440699 HQ401119 JQ889181 JQ889182 JX440540 GU366100 JX440596 JX440597 JX440598 JX440648 JX440649 JX440650 JX440700 JX440701 JQ945302 JX440541 JX440542 JX440599 JX440600 JX440651 JX440652 JQ945303 JX440702 JX440543 JX440544 HQ401198 EU268350 JX440601 JX440653 JX440703 JX440602 EU268319 EU268289 HQ401190 JX440545 JX440603 JX440654 JX440655 JX440704 JX440705 JX440604 JX440656 JX440706 JX440605 JX440606 JX440607 JX440608 HQ426185 JX440657 JX440658 JX440659 HQ426274 JX440707 JX440708 JX440709 JX440609 JX440660 JQ945304 JX440610 JX440661 JX440710 JX440711 JX440552 JX440553 GU550740 JX440554 JX440555 JX440611 JX440612 JX440662 JX440663 JX440613 JX440614 JX440664 JX440713 JX440714 JX440556 JX440615 JX440665 JX440715 GU550732 GQ257747 JX440546 HQ401150 JX440547 JX440548 JQ889185 JX440549 GU550727 JX440550 HQ401163 JX4405510 P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 CAS 226141 JX440712 995 (continued on next page) 996 JX440726 JX440725 JX440673 JX440622 JX440674 JX440724 JX440672 JX440563 HQ401203 JX440564 JX440565 JX440723 JX440671 JX440562 JX440623 JX440624 JX440722 JX440561 JX440720 JX440721 JX440620 JX440621 JX440669 JX440670 JX440716 JX440717 JX440718 JX440719 JX440560 JQ889189 GU366083 PDC JX440666 JX440667 JX440668 JX440616 JX440617 JX440618 JX440619 MXRA5 JQ889177 JX440557 JX440558 JX440559 implemented in PAUP" v4.0 (Swofford, 2002). A thousand bootstrap replicates for each heuristic search were run with ten random sequence replicates using TBR branch swapping. The 1000 bootstrap replicates were summarized as a strict consensus tree. For ML and Bayesian analyses, the data were divided into 13 partitions, 12 corresponding to each codon position of the protein-coding genes and the 13th grouping the tRNA sequences. A 5-partition scheme dividing the analyses among genes was also employed; the resulting ML and Bayesian trees had no significant differences from the 13-partition scheme, so we report only the results from the 13-partition scheme. Partitioned ML analyses were performed using RAxML HPC v7.2.3 (Stamatakis, 2006) on the concatenated dataset. Best fit evolutionary models were estimated in ModelTest v3.7 (Posada and Crandall, 1998) under the Akaike information criterion (Table 3). The analyses were performed using the more complex model (GTR + I + C) applied to all partitions due to computer programming limitations (see Table 3 for selected models). Maximum likelihood inferences were performed for 200 replicates and each inference was initiated with a random starting tree. Gaps were treated as missing data and clade confidence was assessed using 1000 bootstrap pseudoreplicates employing the rapid hill-climbing algorithm (Stamatakis et al., 2008). Partitioned Bayesian analyses were carried out in MrBayes v3.1.2 (Ronquist and Huelsenbeck, 2003) using default priors, with models of nucleotide substitution determined in ModelTest v3.7 (Posada and Crandall, 1998) (Table 3). Two simultaneous parallel runs were performed with eight chains per run, seven hot and one cold. The analysis was run for 20,000,000 generations and sampled every 2000 generations, by which time the chains had long since reached a stationary position and the average standard deviation split frequency fallen below 0.01. The program Are We There Yet? (AWTY) (Nylander et al., 2008) was employed to plot the log likelihood scores against the number of generations to assess convergence and to determine the appropriate number of burn-in trees. We conservatively discarded the first 25% of the trees as burn-in. Nodal support of 0.95 or higher was considered strongly supported. In addition to analyzing the complete dataset, we also performed separate ML analyses of each individual locus to ensure that there were no strongly conflicting patterns among the loci. These analyses were performed in RAxML 7.2.3 under conditions similar to the combined analysis: data were partitioned by codon position (for mitochondrial data, tRNAs constituted a fourth partition), the GTR + I + C model was employed, each initial analysis was repeated for 200 replicates, and 1000 rapid bootstraps were used to assess branch support. Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus sp. ‘‘Timor’’ Cyrtodactylus sworderi Cyrtodactylus tautbatorum Cyrtodactylus tibetanus Cyrtodactylus tigroides Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus Cyrtodactylus LSUHC 8900 LSUHC 6348 BPBM 23317 CAS 210205 USNM 579045 LSUHC 7685 KU 309319 MVZ 233251 IRSNB 2380 LSUHC 6251 n/a CJS 833 USNMFS 36138 ZRC 2.4851 tiomanensis tripartitus tuberculatus tuberculatus yoshii Species 2.3. Divergence timing and ancestral area analyses ID number Table 1 (continued) semenanjungensis seribuatensis sermowaiensis slowinskii ND2 GenBank accession numbers Locality West Malaysia, Johor, Gunung Panti FR, Bunker Trail West Malaysia, Johor, Pulau Mentigi Papua New Guinea, West Sepik Prov., Parkop, Toricelli Mts. Myanmar, Sagaing Division, Alaungdaw Kathapa National Park, Sunthaik Chaung (tributary to Hkaungdin Chaung) Timor L’Est, Manufahi District, Same, Trilolo River Malaysia, Johor, Endau-Rompin, Peta, Sungai Kawal Philippines, Palawan Island, Palawan Province, Municipality of Brooke’s Point China, Tibet Autonomous Region, Lhasa, 3 km WNW of Potala Palace Thailand, Kanchanaburi Province, Sai-Yok District, Ban Tha Sao West Malaysia, Pahang, Pulau Tioman, Tekek-Juara Trail Papua New Guinea, Misima Island Australia, Northeast Queensland Australia, Northeast Queensland Malaysia, Sabah, Poring, Sungai Kipungit trail RAG1 P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 A timescale of evolution in Cyrtodactylus was estimated in BEAST 1.6.1 (Drummond and Rambaut, 2007). As in the preceding analyses, two separate partitioning schemes were employed. However, the 13-partition analysis was terminated after 150 million generations after failing to reach convergence. Therefore, we report results of the 5-partition analysis. The analysis used a random starting tree, and employed Yule tree priors and a relaxed uncorrelated lognormal clock. The analysis was run for 300 million generations, sampling every 10,000 generations, with the first ten percent of generations discarded as burn-in. Estimated sample sizes (>200 for all parameters) were consulted in Tracer 1.5 (Rambaut and Drummond, 2007) to ensure adequate chain length, and 95% highest posterior densities were calculated to provide credible ranges of nodal divergence dates. Three previously-used calibrations (Heinicke et al., 2011) were employed to date the tree. The divergence between S. roosevelti P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 997 Table 2 Primers used for PCR amplification and sequencing. Gene Primer name Primer reference Sequence ND2 L4437b L5002 PHOF1 PHOR1 R13 R18 RAG1F700 RAG1R700 MXRA5F2 MXRA5R2 Macey and Schulte (1999) Macey and Schulte (1999) Bauer et al. (2007) Bauer et al. (2007) Groth and Barrowclough (1999) Groth and Barrowclough (1999) Bauer et al. (2007) Bauer et al. (2007) Portik et al. (2012) Portik et al. (2012) 50 -AAGCAGTTGGGCCCATACC-30 50 -AACCAAACCCAACTACGAAAAAT-30 50 -AGATGAGCATGCAGGAGTATGA-30 50 -TCCACATCCACAGCAAAAAACTCCT-30 50 -TCTGAATGGAAATTCAAGCTGTT-30 50 -GATGCTGCCTCGGTCGGCCACCTTT-30 50 -GGAGACATGGACACAATCCATCCTAC-30 50 -TTTGTACTGAGATGGATCTTTTTGCA-30 50 -KGCTGAGCCTKCCTGGGTGA-30 50 -YCTMCGGCCYTCTGCAACATTK-30 PDC RAG1 MXRA5 Table 3 Best-fit models for data partitions as determined by AIC, and similar models chosen for phylogenetic analyses. Gene Model selected Model applied ND2 1st pos 2nd pos 3rd pos tRNAs GTR + I + C TVM + I + C GTR + C HKY + C GTR + I + C GTR + I + C GTR + C HKY + C PDC 1st pos 2nd pos 3rd pos K81uf + C HKY + C TIMef + C GTR + C HKY + C GTR + C RAG1 1st pos 2nd pos 3rd pos HKY + C TRN + C K81uf + C HKY + C GTR + C GTR + C MXRA5 1st pos 2nd pos 3rd pos TrN + I + C TVM + I HKY + C GTR + I + C GTR + I HKY + C and S. torrei was calibrated (exponential, mean = 3, offset = 15) based on an amber-preserved fossil Sphaerodactylus from Hispaniola dated 15–20 Ma (Iturralde-Vinent and MacPhee, 1996). The divergence between Oedura and Woodworthia was calibrated (exponential, mean = 17, offset = 16) based on fossil New Zealand ‘‘Hoplodactylus’’ dated to 16–19 Ma (Lee et al., 2008). The divergence between Pygopus and Lialis was calibrated (exponential, mean = 10, offset = 20) based on fossil Pygopus dated to 20– 22 Ma (Hutchinson, 1998). Root height was calibrated (normal, mean = 200, S.D. = 13) based on estimated times of divergence of gekkotans from other squamates (Hugall et al., 2007; Jonniaux and Kumazawa, 2008; Vidal and Hedges, 2005). When the final timetree was obtained, ancestral biogeographic regions occupied by Cyrtodactylus were estimated in Mesquite 2.74 (Maddison and Waddison, 2011) under both MP and ML (Mk1 model) criteria. Areas were treated as unordered categorical variables, dividing the range of Cyrtodactylus into eight regions: Tibet, India/Sri Lanka, West Indochina, Thailand, Indochina, Sunda/Wallacea, Philippines, and Papua. Regions correspond to Cyrtodactylus faunal breaks rather than political borders, in some cases enlarged from traditional definitions, and are defined as follows. Tibet includes all areas to the north of the Himalayas inhabited by Cyrtodactylus. India/Sri Lanka includes those parts of India south of the Indo-Gangetic plain as well as the island of Sri Lanka. West Indochina largely corresponds to Assam and Myanmar, and is defined as the part of mainland Southeast Asia west of the Salween River, as well as the southern slopes of the Himalayas. Central Indochina largely corresponds to Thailand, and is defined as the region east of the Salween River, north of the Isthmus of Kra, and west of the Mekong River, except that those parts of southern Vietnam west of the Mekong River (the greater Mekong Delta region) are excluded. Eastern Indochina is defined as all of Vietnam plus those parts of Cambodia and Laos east of the Mekong River. The Sunda/Wallacea region includes both Sundaland (peninsular Malaysia and islands of the Malay Archipelago east to Wallace’s Line) and Wallacea (islands between Wallace’s Line and Lydekker’s Line). The Philippines include the entire Philippine Archipelago. The Papuan Region includes New Guinea, adjacent Indonesian islands east of Lydekker’s Line, those parts of Melanesia inhabited by Cyrtodactylus (Bismarck Archipelago, Solomon Islands), and northeast Queensland, Australia. Several sampled taxa occur in more than one region as defined above; these were coded as occurring in C. Indochina (intermedius, jarujini, oldhami, peguensis) and E. Indochina (interdigitalis), respectively, based on main areas of occurrence plus distributions of their closest relatives. 3. Results The final combined dataset includes 3786 bp, of which 1808 sites are variable and 1251 are parsimony-informative. Bayesian, ML, and MP analyses of the combined dataset recover highly concordant phylogenies (Fig. 2). Our analyses moderately support (Bayesian PP/ML bootstrap/MP bootstrap = 0.74/78/37) the monophyly of Cyrtodactylus with one exception – the peninsular India/ Sri Lanka endemic genus Geckoella is embedded within Cyrtodactylus. None of the included Palearctic bent-toed gecko genera that have long been associated and confused with Cyrtodactylus, especially Cyrtopodion and Mediodactylus, are part of the Cyrtodactylus + Geckoella clade (hereafter referred to Cyrtodactylus sensu lato). However, the species C. tibetanus, which occurs north of the Himalayas and has occasionally been considered allied to some Palearctic bent-toed geckos (e.g. Szczerbak and Golubev, 1986; Szcerback and Golubev, 1996), is confirmed as a member of Cyrtodactylus (Shi and Zhao, 2010). Basal divergences divide Cyrtodactylus sensu lato into three well-supported monophyletic, geographically-circumscribed groupings: (1) C. tibetanus (Clade A in Fig. 2) of the Tibetan Plateau and the northern slopes of the Himalayas; (2) a ‘‘Myanmar Clade’’ (Clade B in Fig. 2) that includes all ten sampled Cyrtodactylus species of diverse habits, size, and body form from Myanmar and the southern flank of the Himalayas, and probably additional unsampled Cyrtodactylus from this region; and (3) a clade that includes all Geckoella plus Cyrtodactylus from Thailand, Indochina, the Malay Peninsula, Indonesia, the Philippines, Australia, and Melanesia, representing about 134 species (Clades C–M in Fig. 2). There is strong support from our analyses that C. tibetanus is outside all remaining Cyrtodactylus + Geckoella (Bayesian PP/ML bootstrap/ MP bootstrap = 1.00/98/78). Within the large third grouping, several additional geographically and morphologically cohesive clades can be identified. 998 P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 Fig. 2. Maximum likelihood phylogeny of Cyrtodactylus, under the 13-partition scheme (#ln L 71587.112627). Nodes supported by all analyses (Bayesian PP > 0.95, ML and MP bootstrap > 70) are indicated by stars. For nodes not supported by all analyses, support from individual analyses is indicated by black (Bayesian), gray (ML), or white (MP) circles, respectively, or support values are reported when less than 0.95 (Bayesian PP) or 70 (ML and MP bootstrap). P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 Geckoella (Clade D) is recovered as monophyletic, but with poor support (Bayesian PP/ML bootstrap/MP bootstrap = 0.9/29/–). A significantly-supported grouping includes all sampled largebodied species from New Guinea, adjacent island groups in Indonesia and Melanesia, and the Cape York Peninsula of Australia (‘‘Papua Clade’’; Clade I). A large, well-supported group of mainly medium-sized Cyrtodactylus (Clades J–M) includes species ranging from mainland Southeast Asia to the Philippines and northwestern Australia, including one subset of species in northern Borneo and the Philippines (‘‘Philippine Clade’’; Clade M) and another subset of species from the Malay Peninsula, Greater Sundas, Sulawesi, the Lesser Sundas, and the northwestern coast of Australia (‘‘Sunda Clade’’; Clade J). Separate analyses of individual loci show little significant conflict with one another or the phylogenies generated using the combined dataset, suggesting that lineage sorting or other coalescent processes have no effect on phylogeny estimates within Cyrtodactylus sensu lato. Of the clades described in the preceding paragraphs, most are recovered by either three or all four single-locus analyses. The only exceptions are Geckoella (poor support in combined analysis; not recovered in any single-locus analysis) and Clade J (not recovered in the PDC or ND2 analyses), though all Clade J species are still recovered as relatively close relatives. Additionally, in comparing among the four single-locus trees, there are only four instances of conflict among loci, where conflict is defined as a clade receiving ML bootstrap support >70 in one single-locus analysis being incompatible with a clade receiving ML bootstrap support >70 in another single-locus analysis. Specifically, in one analysis C. pantiensis is recovered as most closely related to C. semenanjungensis (PDC, bootstrap support = 87), rather than C. tiomanensis (other loci, bootstrap support = 81–90); in one analysis C. novaeguineae is recovered as most closely related to C. loriae (PDC, bootstrap support = 83), rather than Queensland Cyrtodactylus (other loci, bootstrap support = 86–89); in one analysis C. ayeyarwadyensis is outside a brevidactylus/chrysopylos/gansi clade (ND2, bootstrap support = 88), rather than within this clade (other loci, bootstrap support = 71–75); in one analysis C. darmandvillei is recovered as most closely related to C. batucolus plus C. seribuatensis (ND2, bootstrap support = 76), rather than C. jellesmae (MXRA5, bootstrap support = 89). In each of these cases, conflicts are within geographically-coherent clades and do not affect biogeographic interpretations. The Bayesian relaxed-clock timing analysis also recovered a phylogeny that is largely concordant with the combined ML, MP, and BI phylogenies, differing in branching pattern only at a couple poorly-supported short internal nodes. The analysis shows that Cyrtodactylus diverged from Hemidactylus sometime near the Cretaceous–Paleogene boundary, and that Cyrtodactylus sensu lato has diversified throughout the Cenozoic Era (Fig. 3). Ancestral biogeographic analyses depict a generally southeastward pattern of colonization, with Cyrtodactylus originating in the Palearctic and sequentially colonizing the W. Indochina region, the C. Indochina region, and the E. Indochina and Sunda/Wallacea regions; radiations in Wallacea, the Philippines, and the Papua region (including Melanesia and northeastern Queensland) all trace to Sundaland ancestors. The divergence between the Palearctic species C. tibetanus and other Cyrtodactylus occurred approximately 52 (65–40) Ma (Fig. 3), approximately contemporaneous with the timing of the collision of the Indian and Eurasian tectonic plates (Rowley, 1996). The Myanmar Clade diversified starting 32 (39–23) Ma, the divergence between the two sampled Geckoella occurred 31 (38–24) Ma, and the earliest divergences in the Papuan, Sunda, and Philippine Clades were approximately contemporaneous, about 22 (28–15) Ma. The Australian species C. kimberleyensis and those in the Queensland radiation (C. tuberculatus, C. hoskini, C. mcdonaldi, C. adorus, C. pronarus) diverged from their closest 999 relatives in Timor and New Guinea approximately 3 (5–2) Ma and 13 (18–9) Ma, respectively. 4. Discussion 4.1. Phylogeny Our results support the general supposition that geographically coherent groups of Cyrtodactylus also represent monophyletic groups. The basalmost lineage is represented in this study by only C. tibetanus. It is possible that the other Tibetan/northern Himalayan species (e.g. C. medogense, C. zhaoermii; Shi and Zhao, 2010) are also members of this clade. Members of this group have been problematic with respect to generic allocation and have variously been placed in Tenuidactylus (Szczerbak and Golubev, 1986; Szcerback and Golubev, 1996), Cyrtopodion (Zhao and Li, 1987), and Siwaligekko (Khan, 2003) and some species possess characteristics that are distinctive relative most other Cyrtodactylus, including small body size, a clearly segmented, strongly tuberculate tail (in some species), and large dorsal tubercles. It is possible that some unsampled Himalayan geckos currently assigned to the genera Altigekko or Indogekko also belong to this group, but each of these have putatively diagnostic morphological differences from Cyrtodactylus (Khan, 2003). Monophyly of the Myanmar Clade is strongly supported. Sampling in this clade is especially good and even the most aberrant member of this geographic region, C. brevidactylus—originally considered as possibly allied to Geckoella based on morphological grounds (Bauer, 2002)—is embedded well within this clade. One species included in our phylogeny, C. fasciolatus, belongs to this group as well and extends the range of the group west along the southern flanks of the Himalayas to northwest India. At least two species, C. oldhami and C. peguensis, occur peripherally in Myanmar but are not members of this clade. The ranges of both species are east of the Ayeyarwady and Salween Basins and extend deep into Thailand; their relationships are, not surprisingly, with Thai species. The remaining large clade of geckos includes many other geographic groupings, some of which are exclusive. For example, all Philippine Cyrtodactylus fall in a single clade (intermixed with Borneo species) as do the sampled Geckoella, and all of the Papuan and Queensland species (although at least one other Papuan species is not a member of this clade, Oliver et al., pers. comm.). Central Indochina, Eastern Indochina, and the Sunda region all harbor multiple lineages. One Central Indochinese clade (Clade C: C. jarujini, C. angularis, C. chanhomeae) is sister to all remaining taxa, but another C. Indochinese clade occurs elsewhere in the tree (Clade H: C. tigroides, C. oldhami, C. peguensis), and C. interdigitalis is sister to the Sundaland species C. elok (Clade F). Eastern Indochinese Cyrtodactylus are likewise distributed in multiple clades (Clades E, F, and L, as well as C. irregularis, which is in Clade K (Fig. 2) but apparently not closely related to ‘‘core’’ Clade K), each deeply divergent from their respective sister taxa. Borneo supports two lineages, with three species clustering with the Philippine taxa (Clade M), and three others appearing elsewhere (Clade K). Sri Lanka is represented in or sample only by one species of Geckoella, with its sister taxa in peninsular India. However, Sri Lanka supports five species in the C. fraenatus group, which on morphological grounds appears to be a distinct lineage, perhaps allied to the larger-bodied Myanmar taxa. The largest clade of Sunda/Wallacea species (Clade J) is that sister to the main Vietnamese + Borneo/Philippine clades. This includes a subgroup endemic to the Malay Peninsula and its offshore islands. This in turn is sister to C. marmoratus from Bali and Java, and these are sister to a subgroup including two Malaysian species, one from Sulawesi, two from the Lesser Sundas (including Timor) and a newly discovered species from East Montalivet Island off 1000 P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 Hi ma la y as Papuan region Asia Phillippine region Sunda/Wallacea region E. Indochina region Pacific Ocean India/Sri Lanka region C. Indochina region W. Indochina region Tibet region Indian Ocean H e C. mid C. tibeacty C. slowtanulus C. rus ins s s k C. fasc elli ii C. chr iola y C. gan soptus C. aye si ylos y k C. ha arw s C. bre ien ady v s e C. fea ida is nsi e ct a yl s C. nn u C. jaruand s a C. ang jini lei u C. cha lar n C. dec ho is c m C. trie ane ea e d C. inte ra nsis C. honrme d C. intetree ius n C. elo rdig sis k ita C. pul lis c C. ma hel cr lu t i o s C. gr tu o C. old ide ber h s c C. peg am ula tu u i s C. lori ens a C. trip e is a C. epi rtit r C. lou otic us i u C. salosiad s C. klu monensi ge e s r C. ob i ns u i C. ser stu s m s C. nov ow a a C. ado egu ien s C. pro rus ine is a n C. tub aru e e C. hos rcu s k l C. mcd ini atu s C. irre ona g C. cf. ulaldi c r C. par ond is a C. gris doxoren C. eisemer us sis i C. ma nma r C. “m mor nae ar a q C. ua mo tus C. swodriv ratu i C. semrder rga s” t C. tiomena i us n C. pan ane jun g C. dar tien nsis ens m s is C. jell an is es dv s C. er ma il i l C. bat bua e ei u t C. sp. col ens “T us is k C. im im C. ma berl or” l e C. cav aya yen e n C. con rni us sis s c C. pub obr olu s C. annisul inus c C. jamulat us u b t C. au an s t g C. bal bato an u C. yos ens rum h i C. aur ii s e C. red nsi i s C. agu mic ph sa ulu ill nen s ip s in is ni cu s Australia A B C D E F G H I K L J K M 0 Ma Neogene Q Paleogene 25 Ma 50 Ma K 75 Ma Fig. 3. Historical biogeography of Cyrtodactylus. The phylogeny is a Bayesian timetree of Cyrtodactylus, with 95% HPD intervals depicted at key nodes. Branches are colored according to the results of the ML ancestral area reconstruction. Branches where no one inferred area received at least twice as much support as the next most likely area were considered equivocal and are colored black. Other colors correspond to the map, which depicts major distributional regions of Cyrtodactylus. the coast of the Kimberley region of Western Australia (Bauer and Doughty, 2012). Remaining Sundaland species (Clade G) are more closely related to the Papuan Clade and some Indochinese species. 4.2. Biogeography Our results depicting a Palearctic origin of Cyrtodactylus are in general agreement with biogeographic interpretations that have been made under the assumption that Himalayan bent-toed geckos (including C. tibetanus and the unsampled genera Siwaligekko, Altigekko, and Indogekko) represent a ‘‘transition’’ between Palearctic naked-toed geckos such as Cyrtopodion on one hand, and more typical Indo-Australian Cyrtodactylus on the other (e.g., Khan, 2009; Shi and Zhao, 2010; Szcerback and Golubev, 1996). Not sampling Siwaligekko, Altigekko, or Indogekko has little impact on this biogeographic interpretation, as on morphological grounds they are certainly not embedded in core Cyrtodactylus (Clades B–M), and any placement for these genera still would recover a Palearctic origin for Cyrtodactylus. The subsequent spread of Cyrtodactylus to the east superficially resembles the pattern evident in other gekkonid geckos with similar distributions, such as Gehyra, which originated in Asia and colonized the southwest Pacific (Heinicke et al., 2011). However, Cyrtodactylus is significantly more diverse in both mainland Asia and Indonesia, and species of Cyrtodactylus tend to have limited ranges, whereas most Asian and Indonesian Gehyra species are more widespread. P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 A number of fine-scale biogeographic patterns within Cyrtodactylus deserve comment. Our data cannot conclusively resolve the colonization route of the peninsular Indian/Sri Lankan species in Geckoella, as basal relationships within core Cyrtodactylus are not fully resolved. The time of divergence post-dates the Indian/ Eurasian plate collision (Rowley, 1996), so an overland route from Myanmar was available. However, Geckoella is in a more derived position in the phylogeny than are the Myanmar taxa, and the ancestral area analysis does suggest it is most likely that the ancestor of Geckoella + Cyrtodactylus Clades E–M occupied Central Indochina. Thus, Geckoella may have originated via a cross-Bay of Bengal dispersal event, which would also explain the range gap between Geckoella and other Cyrtodactylus. In mainland Southeast Asia, the strong geographic segregation of separate Cyrtodactylus clades in Myanmar, Thailand, Eastern Indochina, and the Malay Peninsula hints at the importance of long-term geographic barriers in promoting regional endemism (see Fig. 1 for locations of potential barriers). The Isthmus of Kra, separating the Thailand region from the Sunda region (which includes the Malay Peninsula), has long been recognized as such a barrier both currently as the location of an abrupt ecological transition from evergreen rainforest to tropical deciduous forest and historically as a zone affected by marine transgressions through much of the Paleogene (Hughes et al., 2003; Woodruff, 2003; Woodruff and Turner, 2009). These transgressions could have isolated Thai and Sunda Cyrtodactylus radiations from one another as they diversified in the mid-Cenozoic. Members of the Myanmar clade occur west of the Ayeyarwady River or in isolated highland regions between the Ayeyarwady and Salween Rivers. These river valleys have likely acted as dispersal barriers due to a lack of suitable habitat (most mainland Cyrtodactylus prefer karstic or otherwise rocky terrain). Likewise, the Mekong valley may serve as a barrier between Thai and Indochinese Cyrtodactylus. An analysis of the entire Indochinese herpetofauna suggests the Mekong is not a major herpetofaunal biogeographic barrier (Bain and Hurley, 2011). However, it does serve as a barrier for other groups (Meijaard and Groves, 2006) and a preference for rocky terrain separates Cyrtodactylus from many other reptile and amphibian groups – the only Cyrtodactylus known from the Mekong valley occur in isolated rocky portions of the Mekong Delta (Bain and Hurley, 2011). Our sampling of the relatively species-rich Cyrtodactylus faunas of Thailand, Cambodia, Laos, and Vietnam is quite limited, however, so we cannot discount the possibility that there is extensive geographic overlap between mainland Southeast Asian Cyrtodactylus clades that our dataset did not capture. In the Malay Archipelago, our analyses clearly indicate that all Philippine Cyrtodactylus species are closely related to species in Borneo. Unlike previous studies that have recovered a monophyletic Philippine radiation (Siler et al., 2010; Welton et al., 2010a, 2010b), our tree indicates two Borneo to Philippines colonization events in the early Neogene, about 10–20 Ma. However, support values for some of the branches in the larger Philippine/Borneo group are not significant, so it is possible that a larger data set would recover a monophyletic set of Philippine species. Within the Philippines, there is a pattern of south to north colonization, in agreement with previous studies. This overall pattern is in contrast to that inferred in another major Philippine gekkonid radiation, Gekko. Philippine Gekko species are monophyletic, but do not have close relatives in Borneo (Rösler et al., 2011). Instead, based on molecular clock dates, Philippine Gekko most likely rafted on Palawan after rifting from the Asian mainland about 30 Ma (Siler et al., 2012). A similar scenario has been proposed in frogs, with Borneo being colonized from Palawan (Blackburn et al., 2010). While such a scenario is not compatible with the topology and divergence dates we infer for Cyrtodactylus, we cannot wholly discount this possibility based on low support values for some of 1001 these branches. One species in the Philippine/Borneo clade, C. aurensis, occurs in the Seribuat Archipelago east of peninsular Malaysia. Two other sampled species, C. seribuatensis and C. tiomanensis, are also endemic to this archipelago. Interestingly, none of these species are closely related to one another, clearly indicating that the South China Sea has not been a major barrier to dispersal in Cyrtodactylus. In Australia, the close relationship between C. kimberleyensis and an undescribed Timorese species suggests that Western Australia was colonized over water from the Lesser Sundas. Unsampled species from the Lesser Sundas (including C. laevigatus, C. wetariensis and several undescribed species) are similarly small (maximum 73 mm SVL) and similar in general morphology, suggesting that there is a regional radiation. Not surprisingly, given the single species in Western Australia and its very marginal distribution, this colonization of Australia is much younger than that that came via New Guinea and is represented by a group of Queensland taxa (including C. tuberculatus), most only recently recognized (Shea et al., 2012). Each of these Australian colonization events may have been facilitated by expansion of exposed land during periods of low sea level. Another Lesser Sunda species, C. darmandvillei, is related to the single species sampled from Sulawesi, C. jellesmae. Sulawesi supports four additional named species and many more remain to be described (D. Iskandar, pers. comm.). The composite geological nature of Sulawesi raises the possibility that it may have a compound fauna of bent-toed geckos, and several animal clades are known to have colonized Sulawesi multiple times (Stelbrink et al., 2012). Based on morphological traits, it has already been suggested that the Sulawesi Cyrtodactylus fauna is comprised of multiple distinct lineages (Iskandar et al., 2011). Divergence times obtained in this study are quite similar to dates recovered for some Cyrtodactylus species by Siler et al. (2012), even though our studies differ in calibration choice (Siler et al. calibrated the divergence between Cyrtodactylus and Gekko) and overall taxon sampling (Siler et al. included six Cyrtodactylus species). For example, we estimate that the lineage leading to C. philippinicus diverged from that leading to C. baluensis 18 (23–13) Ma, compared to an estimate of 20 (25–15) Ma in Siler et al. (2012). Estimated mean rates of molecular evolution are also similar to those estimated for other gekkonids. For example, the estimated mean rate of molecular evolution for RAG1 in our study was 7.1 ! 10#4 substitutions per site per million years, compared to 6.5 ! 10#4 estimated for the genus Gehyra in a previous study (Heinicke et al., 2011). 4.3. Taxonomy It may be argued that Cyrtodactylus is a large and still growing genus that has already become unwieldy to systematists. This phylogenetic analysis then could provide an opportunity to dismantle the genus in a way consistent with the well-supported monophyletic groups hypothesized. It is clear that Geckoella is embedded within other Cyrtodactylus and that the recognition of former as a valid genus would render the latter paraphyletic. Geckoella are indeed morphologically distinctive, with small, relatively stout bodies, short tails, and noticeably large dorsal scales, and also geographically separated from most other Cyrtodactylus, being restricted to South Asia. However, C. brevidactylus, a member of the Myanmar clade, is morphologically very similar to Geckoella. To be consistent with our tree topology recognition of Geckoella would imply the recognition of at least four new genera, one for C. tibetanus (clade A), one for the Myanmar clade (Clade B), one for some Thai species (Clade C), and one for Clades J–M. Our results support that C. tibetanus is more closely related to typical Cyrtodactylus than to the Palearctic naked-toed clade. However, we do not know which, if any, other Tibetan/Himalayan geckos might belong in this 1002 P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 group (Khan, 2003, 2009; Shi and Zhao, 2010; Szcerback and Golubev, 1996). Further, the position of C. tibetanus as sister to all other Cyrtodactylus does not receive more than moderate support in any of the analyses. The Myanmar clade receives the highest level of support of any clade in the genus. However, morphologically, the members of this clade span almost the entire range of morphological variation in the genus, including very large species, miniaturized species, short-fingered ground dwellers, and ‘‘average-sized’’ climbing species. Because of these issues we advocate continued recognition of the entire bent-toed clade as a single genus for the time being, although more detailed morphological studies that may reveal unambiguous synapomorphies of each of the major lineages might prompt the future dismantling of Cyrtodactylus. With respect to Geckoella we recommend the use of this name at the subgeneric level to recognize this distinctive monophyletic lineage within Cyrtodactylus. Geckoella has been regularly employed as a subgeneric name in the past, so such a change is minimally disruptive (Bauer, 2002, 2003; Rösler, 2000; Ulber and Gericke, 1988). There are two other available genus-group names currently in the synonymy of Cyrtodactylus: Puellula Blyth 1861 (type species Puellula rubida Blyth 1861) and Quantasia Wells and Wellington 1985 (type species Hoplodactylus tuberculatus Lucas and Frost 1900). As we did not sample C. rubidus, a species endemic to the Andaman Islands and adjacent Cocos Islands in the Bay of Bengal, it is unclear to which clade that name would apply. Quantasia could be used for the Papuan clade (Clade I), but we do not recommend such usage at this time. 4.4. Conclusions The phylogeny we present in this study does offer a guide to search for synapomorphies or potential diagnostic differences among clades. When such diagnostic traits are identified, the generic- and subgeneric-level taxonomy of Cyrtodactylus can be reevaluated. In addition, it provides a scheme for outgroup selection for regionally focused studies and a guide for making appropriate comparisons for new species descriptions. Finally, the phylogeny we present also opens the possibility of analyzing the biogeography, ecomorphology, and other evolutionary aspects of Cyrtodactylus biology in a phylogenetic, hypothesis-driven context. For example, long-limbed, flat-bodied cave-dwelling species occur in multiple places in the phylogeny, and the strong geographic signal in the phylogeny suggests that convergence in other traits could be similarly widespread in Cyrtodactylus. Acknowledgments We thank Rafe Brown, Lee Grismer, Ross Sadlier, Paul Doughty, Jim McGuire, Fred Kraus, Anslem de Silva, Indraneil Das, Jon Boone, Firoz Ahmed, Eric Smith, Mohammed Al-Mutairi, Montri Sumontha, Rod Hitchmough, Jens Vindum, Alan Resetar, and George Zug for the genetic material used in this study. Initial laboratory work was carried out by Eli Greenbaum, Christie Buonpane and Sayantan Biswas. The authors were supported by National Science Foundation Grant DEB 0844523 to AMB and TRJ. References Bain, R.H., Hurley, M.M., 2011. A biogeographic synthesis of the amphibians and reptiles of Indochina. Bull. Am. Mus. Nat. Hist. 360, 1–138. Bauer, A.M., 2002. Two new species of Cyrtodactylus (Squamata: Gekkonidae) from Myanmar. Proc. Calif. Acad. Sci. 53, 75–88. Bauer, A.M., 2003. Descriptions of seven new Cyrtodactylus (Squamata: Gekkonidae) with a key to the species of Myanmar (Burma). Proc. Calif. Acad. Sci. 54, 463– 498. Bauer, A.M., DeSilva, A., Greenbaum, E., Jackman, T.R., 2007. A new species of day gecko from high elevation in Sri Lanka, with a preliminary phylogeny of Sri Lankan Cnemaspis (Reptilia: Squamata: Gekkonidae). Mitt. Mus. Naturk. Berlin Zool. Reihe 83, 22–32. Bauer, A.M., Doughty, P., 2012. A new bent-toed gecko (Squamata: Gekkonidae: Cyrtodactylus), from the Kimberley region, Western Australia. Zootaxa 3187, 32–42. Blackburn, D.C., Bickford, D.P., Diesmos, A.C., Iskandar, D.T., Brown, R.M., 2010. An ancient origin for the enigmatic flat-headed frogs (Bombinatoridae: Barbourula) from the islands of Southeast Asia. PLoS One 5, e12090. Chan, K.O., Norhayati, A., 2010. A new insular species of Cyrtodactylus (Squamata: Gekkonidae) from northeastern Peninsular Malaysia, Malaysia. Zootaxa 2389, 47–56. Drummond, A.J., Ashton, B., Buxton, S., Cheung, M., Cooper, A., Duran, C., Field, M., Heled, J., Kearse, M., Markowitz, S., Moir, R., Stones-Havas, S., Sturrock, S., Thierer, T., Wilson, A., 2011. Geneious v5.4. <http://www.geneious.com>. Drummond, A.J., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by sampling trees BMC. BMC Evol. Biol. 7, 214. Grismer, L.L., Anuar, S., Quah, E., Muin, M.O., Chan, K.O., Grismer, J.L., Norhayati, A., 2010. A new spiny, prehensile-tailed species of Cyrtodactylus (Squamata: Gekkonidae) from Peninsular Malaysia with a preliminary hypothesis of relationships based on morphology. Zootaxa 2625, 40–52. Groth, J.G., Barrowclough, G.F., 1999. Basal divergences in birds and the phylogenetic utility of the nuclear RAG-1 gene. Mol. Phylo. Evol. 12, 115–123. Heinicke, M.P., Greenbaum, E., Jackman, T.R., Bauer, A.M., 2011. Phylogeny of a trans-Wallacean radiation (Squamata, Gekkonidae, Gehyra) supports a single early colonization of Australia. Zool. Scripta 40, 584–602. Hugall, A.F., Foster, R., Lee, M.S.Y., 2007. Calibration choice, rate smoothing, and the pattern of tetrapod diversification according to the long nuclear gene rag-1. Syst. Biol. 56, 543–563. Hughes, J., Round, P., Woodruff, D.S., 2003. The Indochinese–Sundaic faunal transition at the Isthmus of Kra: an analysis of resident forest bird species distributions. J. Biogeogr. 30, 569–580. Hutchinson, M.N., 1998. The first fossil pygopodid (Squamata, Gekkota), and a review of mandibular variation in living species. Mem. Queensland Mus. 41, 355–366. Iskandar, D.T., Rachmansah, A., Umilaela, F., 2011. A new bent-toed gecko of the genus Cyrtodactylus Gray, 1827 (Reptilia, Gekkonidae) from Mount Tompotika, eastern peninsula of Sulawesi, Indonesia. Zootaxa 2838, 65–78. Iturralde-Vinent, M.A., MacPhee, R.D.E., 1996. Age and paleogeographical origin of Dominican amber. Science 273, 1850–1852. Johnson, C.B., Quah, E., Anuar, S., Muin, M.A., Wood, P.L., Grismer, J.L., Greer, L.F., Onn, C.K., Ahmad, N., Bauer, A.M., Grismer, L.L., 2012. Phylogeography, geographic variation, and taxonomy of the Bent-toed Gecko Cyrtodactylus quadrivirgatus Taylor, 1962 from Peninsular Malaysia with the description of a new swamp dwelling species. Zootaxa 3406, 39–58. Jonniaux, P., Kumazawa, Y., 2008. Molecular phylogenetic and dating analyses using mitochondrial DNA sequences of eyelid geckos (Squamata: Eublepharidae). Gene 407, 105–115. Khan, M.S., 2003. Questions of generic designation of angular-toed geckos of Pakistan with descriptions of three new genera (Reptilia: Gekkonidae). J. Nat. Hist. Wildl. 2, 1–9. Khan, M.S., 2009. Intergeneric relations of the angular-toed geckos of circum Western Himalayas (Sauria: Gekkonidae). Pak. J. Zool. 41, 29–34. Kluge, A.G., 2001. Gekkotan lizard taxonomy. Hamadryad 26, 1–209. Kraus, F., 2008. Taxonomic partitioning of Cyrtodactylus louisiadensis (Lacertilia: Gekkonidae) from Papua New Guinea. Zootaxa 1883, 1–27. Krysko, K.L., Rehman, H., Auffenberg, K., 2007. A new species of Cyrtopodion (Gekkonidae: Gekkoninae) from Pakistan. Herpetologica 63, 100–113. Lee, M.S.Y., Hutchinson, M.N., Worthy, T.H., Archer, M., Tennyson, A.J.D., Worthy, J.P., Scofield, R.P., 2008. Miocene skinks and geckos reveal long-term conservatism of New Zealand’s lizard fauna. Biol. Lett. 5, 833–837. Macey, J., Schulte, J., 1999. Molecular phylogenetics, tRNA evolution, and historical biogeography in anguid lizards and related taxonomic families. Mol. Phylo. Evol. 12, 250–272. Maddison, D.R., Maddison, W.P., 2003. MacClade. Sinauer Associates, Sunderland, Massachusetts. Maddison, W.P., Waddison, D.R. 2011. Mesquite: A Modular System for Evolutionary Analysis. Version 2.75. <http://mesquiteproject.org>. Mahony, S., 2009. Taxonomic status of Cyrtodactylus khasiensis tamaiensis (Smith, 1940) and description of a new species allied to C. chrysopylos Bauer, 2003 from Myanmar (Reptilia: Gekkonidae). Hamadryad 34, 62–74. Meijaard, E., Groves, C.P., 2006. The geography of mammals and rivers in mainland Southeast Asia. In: Lehman, S.M., Fleagle, J.G. (Eds.), Primate Biogeography: Prospects and Progress. Springer, New York, pp. 305–329. Ngo, V.T., 2008. Two new cave-dwelling species of Cyrtodactylus Gray (Squamata: Gekkonidae) from southwestern Vietnam. Zootaxa 1909, 37–51. Ngo, V.T., Bauer, A.M., 2008. Descriptions of two new species of Cyrtodactylus Gray, 1827 (Squamata: Gekkonidae) endemic to southern Vietnam. Zootaxa 1715, 27–42. Nguyen, N.S., Orlov, N.L., Darevsky, I.S., 2006. Descriptions of two new Cyrtodactylus Gray, 1827 (Squamata: Sauria: Gekkonidae) from Southern Vietnam. Russ. J. Herpetol. 13, 215–226. Nylander, J., Wilgenbusch, J.C., Warren, D.L., Swofford, D.L., 2008. AWTY (are we there yet?): a system for graphical exploration of MCMC convergence in Bayesian phylogenetics. Bioinformatics 24, 581–583. Oliver, P., Edgar, P., Mumpuni, Iskandar, D.T., Lilley, R., 2009. A new species of benttoed gecko (Cyrtodactylus: Gekkonidae) from Seram Island, Indonesia. Zootaxa 2115, 47–55. P.L. Wood Jr. et al. / Molecular Phylogenetics and Evolution 65 (2012) 992–1003 Oliver, P., Krey, K., Mumpuni, Richards S., 2011. A new species of bent-toed gecko (Cyrtodactylus, Gekkonidae) from the North Papuan Mountains. Zootaxa 2930, 22–32. Portik, D.M., Wood, P.L., Grismer, J.L., Stanley, E.L., Jackman, T.R., 2012. Identification of 104 rapidly-evolving nuclear protein-coding markers for amplification across scaled reptiles using genomic resources. Conserv. Genet. Resour 4, 1–10. Posada, D., Crandall, K.A., 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14, 817–818. Rambaut, A., Drummond, A.J., 2007. Tracer v1.4. <http://beast.bio.edu.ac.uk/Tracer>. Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572–1574. Rösler, H., 2000. Kommentierte Liste der rezent, subrezent und fossil bekannten Geckotaxa (Reptilia: Gekkonomorpha). Gekkota 2, 28–153. Rösler, H., Bauer, A.M., Heinicke, M.P., Greenbaum, E., Jackman, T., Nguyen, Q.T., Ziegler, T., 2011. Phylogeny, taxonomy, and zoogeography of the genus Gekko Lauranti, 1768 with the revalidation of G. reveesii Gray, 1831 (Sauria: Gekkonidae). Zootaxa 2989, 1–50. Rowley, D.B., 1996. Age of initiation of collision between India and Asia: a review of stratigraphic data. Earth Planet Sci. Lett. 145, 1–13. Shea, G., Couper, P., Wilmer, J.W., Amey, A., 2012. Revision of the genus Cyrtodactylus Gray, 1827 (Squamata: Gekkonidae) in Australia. Zootaxa 3146, 1–63. Shi, L., Zhao, H., 2010. A new species of Cyrtodactylus (Reptilia: Squamata: Geckkonidae) from Xizang Autonomous Region, China. Zootaxa 2336, 51–60. Siler, C.D., Oaks, J.R., Esselstyn, J.A., Diesmos, A.C., Brown, R.M., 2010. Phylogeny and biogeography of Philippine bent-toed geckos (Gekkonidae: Cyrtodactylus) contradict a prevailing model of Pleistocene diversification. Mol. Phylo. Evol. 55, 699–710. Siler, C.D., Oaks, J.R., Welton, L.J., Linkem, C.W., Swab, J.C., Diesmos, A.C., Brown, R.M., 2012. Did geckos ride the Palawan raft to the Philippines? J. Biogeogr. 37, 1217–1234. Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688– 2690. Stamatakis, A., Hoover, P., Rougemont, J., 2008. A rapid bootstrap algorithm for the RAxML web servers. Syst. Biol. 57, 758–771. 1003 Stelbrink, B., Albrecht, C., Hall, R., von Rintelen, T., 2012. The biogeography of Sulawesi revisited: is there evidence for a vicariant origin of taxa on Wallace’s ‘‘anomalous island’’? Evolution 66, 2252–2271. Swofford, D.L., 2002. PAUP". Phylogenetic Analysis Using Parsimony ("and Other Methods). Sinauer Associates, Sunderland, Massachusetts. Szczerbak, N.N., Golubev, M.L., 1986. Gecko Fauna of the USSR and Contiguous Regions. Naukova Dumka, Kiev (in Russian). Szcerback, N.N., Golubev, M.L., 1996. Gecko Fauna of the USSR and Contiguous Regions. Society for the Study of Amphibians and Reptiles, St. Louis. Uetz, P., 2012. The Reptile Database. <http://reptile-database.reptarium.cz> (accessed 08.02.12). Ulber, T., Gericke, F., 1988. Zur Problematik der Verwandtschaftsverhältnisse in der Gattung Cyrtodactylus Gray 1827 und Bemerkungen zur Gattung Nactus Kluge 1983 (Reptilia: Sauria: Gekkonidae). – der Versuch einer – auch philosophischen – analyse. Veröff. Naturhist. Mus. Schleusingen 3, 67–74. Vidal, N., Hedges, S.B., 2005. The phylogeny of squamate reptiles (lizards, snakes, and amphisbaenians) inferred from nine nuclear protein-coding genes. C.R. Biol. 328, 1000–1008. Welton, L.J., Siler, C.D., Diesmos, A.C., Brown, R.M., 2009. A new bent-toed gecko (Genus Cyrtodactylus) from southern Palawan Island, Philippines and clarification of the taxonomic status of C. annulatus. Herpetologica 65, 328–343. Welton, L.J., Siler, C.D., Diesmos, A.C., Brown, R.M., 2010a. Phylogeny-based species delimitation of southern Philippines bent-toed geckos and a new species of Cyrtodactylus (Squamata: Gekkonidae) from western Mindanao and the Sulu Archipelago. Zootaxa 2390, 49–68. Welton, L.J., Siler, C.D., Linkem, C.W., Diesmos, A.C., Brown, R.M., 2010b. Philippine bent-toed geckos of the Cyrtodactylus agusanensis complex: multilocus phylogeny, morphological diversity, and descriptions of three new species. Herpetol. Monogr. 24, 55–85. Woodruff, D.S., 2003. Neogene marine transgressions, palaeogeography and biogeographic transitions on the Thai–Malay Peninsula. J. Biogeogr. 30, 551– 567. Woodruff, D.S., Turner, L.M., 2009. The Indochinese–Sundaic faunal transition: an analysis of terrestrial mammal species distributions. J. Biogeogr. 36, 803–821. Zhao, E., Li, S., 1987. The herpetofaunal composition and characteristics of Mount Namjagbarwa, Tibet. Acta Herpetol. Sinica 6, 36–42.