The Great American Biotic Interchange in frogs: Multiple and early colonization of Central America by the South American genus Pristimantis (Anura: Craugastoridae) Nelsy Rocío Pinto-Sánchez a , Roberto Ibáñez b,c , Santiago Madriñán a , Oris I. Sanjur b , Eldredge Bermingham b , Andrew J. Crawford a,b,c,⇑ a Departamento de Ciencias Biológicas, Universidad de los Andes, A.A. 4976, Bogotá, Colombia b Smithsonian Tropical Research Institute, Apartado 0843-03092, Panamá, Republic of Panama c Círculo Herpetológico de Panamá, Apartado 0824-00122, Panamá, Republic of Panama article info Article history: Received 16 August 2011 Revised 20 November 2011 Accepted 24 November 2011 Available online xxxx Keywords: Ancestral area reconstruction DEC Isthmus of Panama Multi-locus molecular phylogenetics Terrarana abstract The completion of the land bridge between North and South America approximately 3.5–3.1 million years ago (Ma) initiated a tremendous biogeographic event called the Great American Biotic Interchange (GABI), described principally from the mammalian fossil record. The history of biotic interchange between continents for taxonomic groups with poor fossil records, however, is not well understood. Molecular and fossil data suggest that a number of plant and animal lineages crossed the Isthmus of Pan- ama well before 3.5 Ma, leading biologists to speculate about trans-oceanic dispersal mechanisms. Here we present a molecular phylogenetic analysis of the frog genus Pristimantis based on 189 individuals of 137 species, including 71 individuals of 31 species from Panama and Colombia. DNA sequence data were obtained from three mitochondrial (COI, 12S, 16S) and two nuclear (RAG-1 and Tyr) genes, for a total of 4074 base pairs. The resulting phylogenetic hypothesis showed statistically significant conflict with most recognized taxonomic groups within Pristimantis, supporting only the rubicundus Species Series, and the Pristimantis myersi and Pristimantis pardalis Species Groups as monophyletic. Inference of ancestral areas based on a likelihood model of geographic range evolution via dispersal, local extinction, and cladogen- esis (DEC) suggested that the colonization of Central America by South American Pristimantis involved at least 11 independent events. Relaxed-clock analyses of divergence times suggested that at least eight of these invasions into Central America took place prior to 4 Ma, mainly in the Miocene. These findings con- tribute to a growing list of molecular-based biogeographic studies presenting apparent temporal conflicts with the traditional GABI model. Ó 2011 Elsevier Inc. All rights reserved. 1. Introduction The completion of the Panamanian Isthmus by 3.5–3.1 mil- lion years ago (Ma) (Coates and Obando, 1996) created a land bridge that precipitated one of the greatest biogeographical events in the hemisphere, the Great American Biotic Interchange or GABI, a model based primarily on the mammalian fossil re- cord (Marshall et al., 1982; Marshall, 1988; Simpson, 1940; Webb, 1978; Webb and Rancy, 1996). The GABI allowed taxa from North and South America to move between continents dur- ing the late Neogene, forever altering the biotic composition of both continents (Morgan, 2005). Before the GABI, the biota of North America had general Holarctic affinities, while South America had existed in ‘‘splendid isolation’’ since the mid-Creta- ceous breakup of Gondwanaland and its separation from Antarc- tica in the late Oligocene (Dacosta and Klicka, 2008; Savage, 1982; Simpson, 1980). Once the Isthmian land bridge was com- plete, the GABI acted as a driver of expansion, extinction, and diversification of lineages on both continents (Marshall et al., 1982). The mode, timing and biotic ramifications of the GABI in non- mammalian taxa are less understood. For lineages with a compara- tively poor fossil record, molecular data have played a crucial role in studying the evolutionary history of biotic exchange between continents. Examples include molecular phylogenetic analyses of 1055-7903/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2011.11.022 ⇑ Corresponding author. Address: Instituto de Genética, Edif. M1-304, Departa- mento de Ciencias Biológicas, Universidad de los Andes, Carrera 1E No. 18A–10, A.A. 4976, Bogotá, Colombia. Fax: +57 1 332 4069. E-mail addresses: [email protected](N.R. Pinto-Sánchez), ibanezr@ si.edu (R. Ibáñez), [email protected](S. Madriñán), [email protected](O.I. Sanjur), [email protected](E. Bermingham), [email protected], (A.J. Crawford). Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx Contents lists available at SciVerse ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of Central America by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
41
Embed
The Great American Biotic Interchange in frogs: Multiple and early ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The Great American Biotic Interchange in frogs: Multiple and early colonizationof Central America by the South American genus Pristimantis(Anura: Craugastoridae)
Nelsy Rocío Pinto-Sánchez a, Roberto Ibáñez b,c, Santiago Madriñán a, Oris I. Sanjur b,Eldredge Berminghamb, Andrew J. Crawford a,b,c,!aDepartamento de Ciencias Biológicas, Universidad de los Andes, A.A. 4976, Bogotá, Colombiab Smithsonian Tropical Research Institute, Apartado 0843-03092, Panamá, Republic of PanamacCírculo Herpetológico de Panamá, Apartado 0824-00122, Panamá, Republic of Panama
a r t i c l e i n f o
Article history:Received 16 August 2011Revised 20 November 2011Accepted 24 November 2011Available online xxxx
Keywords:Ancestral area reconstructionDECIsthmus of PanamaMulti-locus molecular phylogeneticsTerrarana
a b s t r a c t
The completion of the land bridge between North and South America approximately 3.5–3.1 million yearsago (Ma) initiated a tremendous biogeographic event called the Great American Biotic Interchange(GABI), described principally from the mammalian fossil record. The history of biotic interchangebetween continents for taxonomic groups with poor fossil records, however, is not well understood.Molecular and fossil data suggest that a number of plant and animal lineages crossed the Isthmus of Pan-ama well before 3.5 Ma, leading biologists to speculate about trans-oceanic dispersal mechanisms. Herewe present a molecular phylogenetic analysis of the frog genus Pristimantis based on 189 individuals of137 species, including 71 individuals of 31 species from Panama and Colombia. DNA sequence data wereobtained from three mitochondrial (COI, 12S, 16S) and two nuclear (RAG-1 and Tyr) genes, for a total of4074 base pairs. The resulting phylogenetic hypothesis showed statistically significant conflict with mostrecognized taxonomic groups within Pristimantis, supporting only the rubicundus Species Series, and thePristimantis myersi and Pristimantis pardalis Species Groups as monophyletic. Inference of ancestral areasbased on a likelihood model of geographic range evolution via dispersal, local extinction, and cladogen-esis (DEC) suggested that the colonization of Central America by South American Pristimantis involved atleast 11 independent events. Relaxed-clock analyses of divergence times suggested that at least eight ofthese invasions into Central America took place prior to 4 Ma, mainly in the Miocene. These findings con-tribute to a growing list of molecular-based biogeographic studies presenting apparent temporal conflictswith the traditional GABI model.
! 2011 Elsevier Inc. All rights reserved.
1. Introduction
The completion of the Panamanian Isthmus by 3.5–3.1 mil-lion years ago (Ma) (Coates and Obando, 1996) created a landbridge that precipitated one of the greatest biogeographicalevents in the hemisphere, the Great American Biotic Interchangeor GABI, a model based primarily on the mammalian fossil re-cord (Marshall et al., 1982; Marshall, 1988; Simpson, 1940;Webb, 1978; Webb and Rancy, 1996). The GABI allowed taxa
from North and South America to move between continents dur-ing the late Neogene, forever altering the biotic composition ofboth continents (Morgan, 2005). Before the GABI, the biota ofNorth America had general Holarctic affinities, while SouthAmerica had existed in ‘‘splendid isolation’’ since the mid-Creta-ceous breakup of Gondwanaland and its separation from Antarc-tica in the late Oligocene (Dacosta and Klicka, 2008; Savage,1982; Simpson, 1980). Once the Isthmian land bridge was com-plete, the GABI acted as a driver of expansion, extinction, anddiversification of lineages on both continents (Marshall et al.,1982).
The mode, timing and biotic ramifications of the GABI in non-mammalian taxa are less understood. For lineages with a compara-tively poor fossil record, molecular data have played a crucial role instudying the evolutionary history of biotic exchange betweencontinents. Examples include molecular phylogenetic analyses of
1055-7903/$ - see front matter ! 2011 Elsevier Inc. All rights reserved.doi:10.1016/j.ympev.2011.11.022
! Corresponding author. Address: Instituto de Genética, Edif. M1-304, Departa-mento de Ciencias Biológicas, Universidad de los Andes, Carrera 1E No. 18A–10, A.A.4976, Bogotá, Colombia. Fax: +57 1 332 4069.
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
bolitoglossine salamanders (Hanken and Wake, 1982), true dart-poison frogs (Maxson and Myers, 1985), viperid snakes (Zamudioand Greene, 1997), primary freshwater fishes (Bermingham andMartin, 1998), beetles (Zeh et al., 2003), túngara frogs (Weigt et al.,2005), procyonids (Koepfli et al., 2007) and birds (Weir et al.,2009). A growing number of molecular studies suggest that animalsand especially plants may have moved between North and SouthAmerica well before the apparent 3.5–3.1 Ma Isthmian closure date(reviewed in Cody et al. (2010)). While the fossil data show that afewmammal lineages crossed early, e.g., raccoons and sloths (Koep-fli et al., 2007;Marshall et al., 1982;Marshall, 1988), the growing listofmolecular-based studies proposing dispersal events >3.5 Ma sug-gests that the GABI wasmore complex than previous appreciated orthat geological models, fossil evidence, phylogenetic sampling and/or molecular evolutionary analyses should be revisited (Daza et al.,2010; Koepfli et al., 2007). With few exceptions (e.g., Hanken andWake, 1982), molecular studies lack data from the part of SouthAmerica closest to the closure, anddivergence times across thepointof contact between continents could be overestimated. Frogs pro-vide a useful model to study the role of the isthmian land bridge inintercontinental dispersal because they are terrestrial, unable tofly and are intolerant of salt water. In this study, we use improvedtaxonomic and geographic sampling of a group of direct-developingfrogs in Colombia andPanamaandcombine thiswithpublisheddatato infer the number and timing of colonization events betweenSouth and Central America.
The genus Pristimantis Jiménez de la Espada 1870 contains448 species (AmphibiaWeb, 2011; Hedges et al., 2008) com-monly called rain frogs, robber frogs or dirt frogs, that are lar-gely restricted to moist, forested habitats and form part of alarger clade of Neotropical direct-developing frogs called Terrar-ana. In Central America, Pristimantis ranges from Panama north-ward to eastern Honduras, and in South America it ranges fromColombia southward through the Andes to Bolivia, and westwardinto Amazonian Brazil, the Guianas and including the LesserAntilles (AmphibiaWeb, 2011). The bulk of species diversitywithin Pristimantis occurs in the Andes of Colombia, Ecuadorand Peru (Lynch and Duellman 1997; Frost, 2009). Hedgeset al. (2008) recognized three subgenera within Pristimantis[Hypodyction, Pristimantis and Yunganastes (Padial et al., 2007)]and 16 phenetic groups (cf. Lynch and Duellman, (1997)) withinthe subgenus Pristimantis (bellona, chalceus, conspicillatus, curti-pes, devillei, frater, galdi, lacrimosus, leptolophus, loustes, myersi,orcesi, orestes, peruvianus, surdus and unistrigatus). Pristimantiswas placed in the family Strabomantidae by Hedges et al.(2008), and in the family Craugastoridae, subfamily Pristimanti-nae, by Pyron and Wiens (2011).
The geographic distribution of Pristimantis suggests that it orig-inated in South America, home to most of its species and relatedgenera (Duellman, 2001; Savage, 2002; Vanzolini and Heyer,1985). Molecular phylogenetic analyses support a South Americanorigin for this genus (Hedges et al., 2008; Heinicke et al., 2007), yetthese studies suffered from poor taxon sampling near the meetingpoint between Central and South America. Thus, the origins of Cen-tral American Pristimantis remain poorly understood. Given ourimproved sampling within Panama and Colombia, we are now ableto ask the following questions: (1) Are the taxonomic groups rec-ognized by Hedges et al. (2008) for Pristimantis monophyletic?(2) Are Central American Pristimantis derived from South America?(3) Howmany times did Pristimantis invade Central or South Amer-ica? (4) Did Pristimantis cross between Central and South Americaprior to the closure of the Isthmus? To answer these questions, weinferred the genealogical and biogeographic history of Pristimantisfrom mitochondrial and nuclear genes, and compared our resultsto geological reconstructions and the biogeographic histories ofco-distributed organisms.
2. Materials and methods
2.1. Taxon sampling
For our molecular phylogenetic analysis of Pristimantis we be-gan with the taxonomic classification of Hedges et al. (2008). Weobtained new DNA sequence data from 71 individuals representing31 species (Appendix A). Additional sequences representing 107species and 109 individuals were downloaded from GenBank (Sup-plemental material Table S1). While many details of terraranidrelationships remain unclear, recent studies suggest that the sistergroup of Pristimantis is a clade containing the South American gen-era Lynchius, Oreobates and Phrynopus, and the predominantlyCaribbean clade, Eleutherodactylidae, is the sister taxon to the restof Terrarana (Hedges et al., 2008; Heinicke et al., 2009; Pyron andWiens, 2011). We therefore included as samples of close relativesof Pristimantis: Lynchius flavomaculatus, Lynchius nebulanastes,Oreobates cruralis, Oreobates saxatilis, Phrynopus auriculatus andPhrynopus bracki. From Craugastorinae [sensu Pyron and Wiens(2011)] we included Craugastor daryi and Craugastor longirostris,and from Eleutherodactylidae we included Diasporus hylaeformisand Diasporus vocator. We rooted our terraranid phylogeny withthe hylids, Agalychnis callidryas from Central America and Litoriacaerulea from Australia.
Specimens were collected in four countries: Colombia, CostaRica, Panama and Peru (Fig. 1). Additional tissue samples werekindly provided by the Círculo Herpetológico de Panamá (CH),the Colección Herpetológica de la Universidad Industrial de Sant-ander, Colombia (UIS-H), Museo de Herpetología de la Universidadde Antioquia, Colombia (MHUA), the Museo de Zoología de la Pon-tificia Universidad Católica del Ecuador, Ecuador (QCAZ), and theAmphibian and Reptile Diversity Research Center at the University
Fig. 1. Map showing collecting localities for Pristimantis. Circles represent localitiesof new data reported for the first time here, and triangles represent localitiescorresponding to data obtained from GenBank. The dotted line represents the limitused by us to differentiate the distribution of species in Central and South America.Darker shading indicates increased elevation.
2 N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
of Texas at Arlington (UTA). Tissues collected in the field were pre-served in 95% ethanol or in a solution of 20% dimethylsulfoxide(DMSO), 0.125 M EDTA and saturated with NaCl (Amos et al.,1992). Most collected specimens were deposited at public researchinstitutions, with voucher numbers for each specimen and Gen-Bank accession numbers for each gene fragment listed in AppendixA.
2.2. Laboratory techniques
We sequenced fragments of the following three mitochondrialgenes: 16S rRNA (16S), 12S rRNA (12S), and the Folmer or ‘‘Barcodeof Life’’ fragment of the cytochrome oxidase sub-unit I (COI) gene(Meyer and Paulay, 2005; Smith et al., 2008). For a subset of sam-ples we also obtained DNA sequence data from exons of two nucle-ar genes: the recombination activating gene 1 (RAG-1) and thetyrosinase precursor gene (Tyr) (Table 1). We chose mitochondrialgenes because of their fast rate of evolution, as compared to nucle-ar genes, in an attempt to resolve recent divergences (e.g., Guayas-amin et al., 2008). The COI gene fragment was chosen also for itsutility in the global DNA barcoding effort (Crawford et al., 2010a;Smith et al., 2008). Tyr and RAG-1 were chosen because they haveproven useful in previous molecular phylogenetic studies of inter-and intra-generic frog diversification (Bossuyt and Milinkovitch,2001; Frost et al., 2006; Heinicke et al., 2007).
Genomic DNA was extracted from liver or thigh muscle tissueusing the DNeasy Blood & Tissue Kit (Qiagen). PCR amplificationof gene fragments was performed in 12.5 ll reactions using0.125 ll Qiagen Taq, 1.25 ll Buffer 10X with 1.5 mM of MgCl2,1.25 ll dNTPs at 2 mM, 0.625 ll each of forward and reverse prim-ers at 10 mM, and 1 ll of extracted DNA (more for low-qualityextractions). Standard reaction conditions were an initial denatur-ation for 5 min at 94 "C followed by 32 cycles of 94 "C for 30 s,annealing at 55 "C (for 16S and 12S), 52 "C (for COI) or 60 "C(RAG-1 and Tyr) for 30 s, and extension at 72 "C for 60 s. Afterwardsa final extension of 72 "C for 7 min was performed. For low-yield-ing samples, the annealing temperature was lowered to 46 "C. PCRproducts were cleaned by gel slicing and agarose digest or by Exo I/SAP digest. For each individual, both heavy and light strands weresequenced directly. PCR primers were used in cycle sequencingreactions with BigDye reaction mix and following a standard cyclesequencing profile of 96 "C for 60 s followed by 30 cycles of 96 "Cfor 10 s, 50 "C for 15 s and 60 "C for 4 min, and ending with 72 "Cfor 7 min. DNA sequencing was performed with an ABI Prism3100 sequencer.
DNA sequence chromatograms were cleaned with Sequencher4.2 (Gene Codes Corporation) and Geneious 3.7.0 (BiomattersLtd.). Sequences were aligned initially using MAFFT6.0 (Katoh
and Toh, 2010) under default parameters. Manual adjustments tomaintain correct reading frame in protein-coding genes (COI,RAG-1 and Tyr) were made using MacClade 4.08 (Maddison andMaddison, 2005). The 16S and 12S genes are conserved mitochon-drial markers but indel mutations are common in variable regionscorresponding to loops in the ribosomal RNA structure. Ribosomalgene alignments were conducted using G-block 0.91b (Castresana,2000) and evaluated by eye. GenBank accession numbers are as fol-lows: JN991416–JN991480 for 16S, JN991481–JN991549 for 12S,JN991345–JN991415 for COI, JQ025165–JQ025214 for RAG-1 andJN991550–JN991598 for Tyr.
Alignments are available at TreeBASE (http://www.tree-base.org) under URL http://purl.org/phylo/treebase/phylows/study/TB2:S11988. All DNA and sample datamay also be found at Barcodeof Life Data Systems (Ratnasingham and Hebert, 2007) underproject code BSMPE.
2.3. Phylogenetic analyses
Prior to concatenated analyses, single gene datasets were in-spected for significant incongruence (Wiens, 1998) by comparingpreliminary neighbor-joining (NJ; Saitou and Nei, 1987) and max-imum parsimony (MP) trees obtained using PAUP! 4.0b10 (Swof-ford, 2002), with preliminary support evaluated by 2000 non-parametric bootstrap pseudo-replicates (Felsenstein, 1985), eachemploying ten replicates of random taxon addition. PreliminaryNJ trees were based on HKY distances (Hasegawa et al., 1985),while MP inference used heuristic searches with 100 random-addi-tion sequence replicates and tree bisection–reconnection (TBR)branch swapping. We did not apply an Incongruence Length Differ-ence test because of potentially inflated Type I error rates (Barkerand Lutzoni, 2002).
Phylogenetic analyses were conducted using MP, maximumlikelihood (ML), and Bayesian methods on individual genes andon concatenated datasets (see below). For MP analyses we per-formed a heuristic search with 10000 replicates of random taxonaddition and TBR branch swapping using PAUP!v4.0b10 availableon the CIPRES portal (Miller et al., 2010). Non-parametric bootstrapvalues were obtained with 5000 replicates, each having ten repli-cates of random taxon addition.
Prior to ML and Bayesian analyses, we used Modeltest 3.7 (Po-sada and Crandall, 1998) and MrModeltest 2.3 (Nylander, 2004)to select the optimal model for each data partition (see below)according to the Akaike information criterion, or AIC (Akaike,1973), which allows one to compare non-nested models and to ac-count for model-selection uncertainty using multi-model inference(Posada and Buckley, 2004). Maximum likelihood analyses wererun in RAxML 7.0.4 (Stamatakis et al., 2008), which uses the model,
Table 1Primers employed in this study for PCR and DNA sequencing.
Gene region Primer name Primer sequence (50–30) Source
Mitochondrial COI LCO-1490 GGTCAACAAATCATAAAGATATTGG Folmer et al. (1994)dgLCO-1490 GGTCAACAAATCATAAAGAYATYGG Meyer et al. (2005)HCO-2198 TAAACTTCAGGGTGACCAAAAAATCA Folmer et al. (1994)dgHCO-2198 TAAACTTCAGGGTGACCAAARAAYCA Meyer et al. (2005)
Mitochondrial 16S Sar-L CGCCTGTTTATCAAAAACAT Palumbi et al. (1991)Sbr-H CCGGTCTGAACTCAGATCACGT Palumbi et al. (1991)
Mitochondrial 12S H10 CACYTTCCRGTRCRYTTACCRTGTTACGACTT Heinicke et al. (2007)L4E TACACATGCAAGTYTCCGC Heinicke et al. (2007)
Nuclear RAG-1 R182 GCCATAACTGCTGGAGCATYAT Heinicke et al. (2007)R270 AGYAGATGTTGCCTGGGTCTTC Heinicke et al. (2007)
Nuclear Tyr Tyr1C GGCAGAGGAWCRTGCCAAGATGT Bossuyt and Milinkovitch (2001)Tyr1G TGCTGGGCRTCTCTCCARTCCCA Bossuyt and Milinkovitch (2001)
N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx 3
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
GTRCAT, as an approximation of the GTR + C model. Node supportwas assessed via 1000 bootstrap replicates.
We conducted Bayesian phylogenetic analyses using MrBayes3.1 (Ronquist and Huelsenbeck, 2003) as implemented in theCIPRES portal under the models of evolution recommended viaMrModeltest. After conducting shorter test runs, we conductedthree parallel runs of the Metropolis coupled Monte Carlo Markovchain (MCMCMC) algorithm for 10 million generations each, sam-pling one tree with associated parameter values per 1000 genera-tions, and employing two heated chains with a 0.05 heatingparameter. Convergence and stationarity of the Markov processwere evaluated via average split frequencies less than 0.05 amongruns and by the stability and adequate ‘‘mixing’’ of sampled log-likelihood values and parameter estimates across generations asvisualized using Tracer 1.3 (Rambaut and Drummond, 2004). Thefirst 1 million generations were discarded as burn-in.
2.3.1. Data partitionsBecause our combined data set comprised one protein-coding
mitochondrial gene (COI), two ribosomal genes with secondarystructure (12S and 16S), and two nuclear genes (RAG-1 and Tyr),application of a single nucleotide substitution model was unlikelyto provide a particularly good fit to the data (Brandley et al., 2005;Nylander et al., 2004). Partitions were chosen a priori based ongene identity (12S, 16S, COI, RAG-1 and Tyr) and codon position.We evaluated three distinct partitioning strategies, including nopartition, a 5-way partition by gene and an 11-way partition bygene and codon position (see Section 3).
We used three statistics to choose the best-fit partitioned modelfor analysis of the combined data: (1) Bayes factors (2lnB10), (2)relative Bayes factors (RBF), and (3) Akaike weights (Aw) (Castoeand Parkinson, 2006; Castoe et al., 2005). Bayes factors were calcu-lated using twice the difference in the marginal model likelihoodsas estimated from the harmonic mean of the sample of posteriortrees (Nylander et al., 2004). Values greater than 10 were consid-ered very strong evidence that the more complex model explainedthe data better (Kass and Raftery, 1995; Nylander et al., 2004). Rel-ative Bayes factors (RBF) (Castoe et al., 2005) were used to quantifythe average impact that each free model parameter had on increas-ing the fit of the model to the data (Castoe and Parkinson, 2006),and were calculated as Bayes factors divided by the difference innumber of free parameters in the two models under consideration.Akaike weights (Aw) measure the support for model i relative tothe model with the lowest AIC value (Castoe et al., 2005), and werecalculated as exp("DAIC/2), where DAIC = AICi "minAIC.
2.3.2. Testing the monophyly of taxonomic groupsWe tested monophyly of the following taxonomic groups with-
in Pristimantis: conscipillatus, curtipes, devillei, frater, lacrimosus,myersi, orcesi, orestes, pardalis, peruvianus, surdus and unistrigatus,following the taxonomy of Hedges et al. (2008). Twelve con-strained tree topologies were constructed with MacClade 4.08(Maddison and Maddison, 2005). For each test, we conducted anew ML tree search constraining a single node such that all sam-pled members of a given taxonomic group under evaluation wereforced to be monophyletic. The significance of the difference inthe sum of site-wise log-likelihoods for all trees was evaluatedby resampling estimated log-likelihoods (RELL bootstrapping) ofsite scores with 1000 replicates, then calculating how far a givenobserved difference was from the mean of the RELL sampling dis-tribution (Shimodaira and Hasegawa, 1999). The constrainedtopology was compared to the unconstrained ML topology usingthe paired-sites test (SH) of Shimodaira and Hasegawa (1999) asimplemented in PAUP!. The SH test is a conservative test of treetopology (Crawford et al., 2007; Felsenstein, 2004).
2.4. Historical biogeography
2.4.1. Divergence timesDivergence times along with phylogenetic relationships were
estimated for the complete data set (202 individuals and4074 bp) using the program BEAST 1.5.4 (Drummond and Ram-baut, 2007), and assuming the 11-way partition scheme (see Sec-tion 3) with a relaxed clock, allowing substitution rates to varyaccording to an uncorrelated Lognormal distribution (Drummondet al., 2006). We assumed a Yule tree prior, i.e., a constant specia-tion rate per lineage (Drummond et al., 2006). To estimate the agesof Central American lineages of Pristimantis we constrained theroot node of our phylogeny along with one node outside of thisgenus (see below). To explore the sensitivity of divergence timeestimates to uncertainty in the root age, we ran five separate anal-yses assuming various published hypotheses for the age of thisnode representing the common ancestor of Terrarana and hylidfrogs (Heinicke et al., 2007; Roelants et al., 2007; Wiens, 2007,2011; Wiens et al., 2011) (Table 6). In all cases the prior distribu-tion for the root age was assigned a normal distribution withapproximately the same mean as the point estimates obtained bythese authors and a standard deviation (SD) that approximatedthe uncertainty around these estimates (see below).
In the first divergence time analysis we assumed a prior meandivergence date of 92.31 Ma with a SD of 20 million years in orderto include the maximum age of 131.2 Ma within 2 SD of the mean(Wiens, 2007). The second, third, and fourth divergence-time anal-yses assumed a mean prior age of 80 Ma, 70 Ma and 50 Ma, respec-tively, with a SD of 10 million years based on the results of Wienset al. (2011), Wiens (2011) and Roelants et al. (2007), respectively.The fifth divergence-time analysis assumed a mean prior root ageof 57 Ma with a SD of 9 million years (Heinicke et al., 2007). Inall analyses, a second calibration interval was applied to the stemage of the Central American genus, Craugastor, applying a mean of42 Ma and a SD of 7 million years, following Heinicke et al. (2007)rather than the older age suggested by Crawford and Smith (2005).Our priors on divergence times therefore represented secondarycalibrations. Details regarding the primary calibrations are foundin the respective references cited above. All other priors were leftto their default values. Parameters were estimated using two inde-pendent runs of 90 million generations each with a burn-in of9 million generations and trees sampled every 10 thousand gener-ations. Convergence was checked in the Tracer 1.5 program, andsummary trees were generated using TreeAnnotator 1.5.4, bothpart of the BEAST package. Using Tracer, we confirmed that ourpost-burnin trees yielded an effective sample size (ESS) of >200for all model parameters, including the ages of all clades of interest(Table 6).
The minimum age of colonization of a new area from a sourceregion may be estimated from the time to the most recent commonancestor (TMRCA) of a clade endemic to the new area, i.e., thecrown age of said clade. The maximum age of colonization maybe estimated as the TMRCA of the endemic clade and its closest rel-ative in the source region, i.e., the stem age of said clade. Here weare interested specifically in minimum ages of Central Americanlineages of a largely South American genus (see below), and focustherefore on estimates and confidence intervals of crown ages.
2.4.2. Ancestral area reconstructionTo investigate the geographic origins of Pristimantis and its sub-
sequent history of dispersal between South and Central America,we reconstructed ancestral areas for Pristimantis using a likeli-hood-based method for inferring geographic-range evolutionthrough dispersal, local extinction and cladogenesis (DEC), asimplemented in the program, Lagrange 2.0.1 (Ree et al., 2005;Ree and Smith, 2008). For this analysis, we assumed the resulting
4 N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
Bayesian consensus phylogeny (see below) but, due to restrictionsin the software implementation, we trimmed the complete treedown to 179 tips by removing potentially redundant conspecificsamples. Analyses were run using default settings with no priorconstraints. In the DEC model, dispersal events cause range expan-sion, local extinction events cause range contraction, and thecumulative probability of each type of event between any twonodes is proportional to the branch length (Clark et al., 2008). Be-cause we were interested only in possible dispersal events be-tween South (SA) and Central America (CA), taxa were codedusing these two discrete character states plus a third characterstate for species or ancestors present in both regions. We also in-ferred ancestral areas with standard parsimony methods usingMesquite 2.74 (Maddison and Maddison, 2009), with orderedstates such that the widespread state was intermediate to SA andCA.
3. Results
3.1. Phylogenetic analysis
The complete data set of five gene fragments contained4074 bp, 612 bp of COI, 1385 bp of 16S, 913 bp of 12S, 633 bp ofRAG-1 and 531 bp of Tyr, obtained from 202 individuals represent-ing 138 species (Table 2). No premature stop codons were detectedin the three protein-coding genes. We observed no significant con-flict among individual gene trees, and we present the combined-data analyses here. ModelTest selected the GTR + I + C model asoptimal for most genes (Table 3). Statistical comparisons of alter-native partitioning schemes gave strong support for the 11-waypartitioned model over the 5-way and unpartitioned alternatives(Table 4). Bayesian inference yielded a consensus tree (Fig. 2) that
was topologically congruent with the ML trees and presented nosignificant conflict with MP inference. MP bootstrap support andBayesian posterior probabilities were largely consistent amongnodes (Fig. 2).
Pristimantis was monophyletic with significant support in MP,ML and Bayesian analyses and was placed as the sister taxon tothe clade comprising Lynchius, Oreobates and Phrynopus (Fig. 2,Supplementary material Fig. S1). Two subgenera of Pristimantis[Hypodictyon and Pristimantis (Hedges et al., 2008)] were not recov-ered as monophyletic. Although we have only one species repre-senting the subgenus Yunganastes (Padial et al., 2007), thissample was placed within the P. peruvianus Species Group of thesubgenus Pristimantis, contra Padial and de la Riva (2009).
Within the subgenus Hypodictyon, the Pristimantis ridens Spe-cies Series sensu Hedges et al. (2008) was not recovered as a mono-phyletic group, while the P. rubicundus Species Series sensuCrawford et al. (2010b) was recovered. Within the subgenus Pristi-mantis, monophyly was rejected for the P. conscipillatus, P. curtipes,P. devillei, P. frater, P. lacrimosus, P. orcesi, P. orestes, P. peruvianus, P.surdus and P. unistrigatus Species Groups sensu Hedges et al. (2008)by the SH test at P < 0.05 (Table 5). Groups not rejected as mono-phyletic included the P. myersi Species Group and the P. pardalisSpecies Group sensu Wang et al. (2008). From the P. chalceus andP. galdi Species Groups only one species was sampled, so theirmonophyly could not be tested with the present data set (Table 5,Supplementary material Table S2).
3.2. Divergence times, ancestral reconstruction and biogeographicalhistory
Divergence-time analyses under our five alternative scenariosfor the root age of the phylogeny gave concordant results regardingthe estimated crown ages of taxa with Central American represen-tatives (Table 6). In the following discussion we cite divergencetime estimates obtained from the relatively young calibrationinterval based on Heinicke et al. (2007). According to this analysisthe genus Pristimantis diverged from other eleutherodactylines inthe Eocene 52 Ma (with 95% credibility interval, CI, of 39–66 Ma)and began radiating 38 Ma (CI: 28–49 Ma) (Fig. 3). Most of the ba-sal splits in Terrarana gave rise to South American taxa, and notsurprisingly the DEC analysis placed the origin of the genus Pristi-mantis in this continent (Fig. 3).
The ancestral-area reconstruction using either parsimony orDEC showed at least 11 separate dispersal events from South
Table 2Number and proportion of invariant, variable but un-informative (singletons), andparsimony informative (PI) sites for each gene region. In each column the number ofsites is given first, with the corresponding proportion in parentheses.
Gene Aligned positions Invariant sites Singleton sites PI sites
Table 3Estimated parameters were calculated using MrModeltest 2.3. (Nylander, 2004) and PAUP! v4.0b10 (Swofford, 2002). I indicates the proportion of invariable sites and the shapeparameter a determines the relative frequency of rates among sites following a C-distribution.
Gene Best-fit model I Shape parameter, a Rate matrix Base frequency
AC AG AT CG CT GT A C G T
COI GTR + I + C 0.4311 0.5459 0.5680 11.2552 0.4570 0.4277 6.3348 1.0000 0.3069 0.3028 0.1010 0.289312S GTR + I + C 0.2627 0.7028 2.6346 9.1316 2.6776 0.4414 20.9307 1.0000 0.3888 0.2278 0.1621 0.221316S GTR + I + C 0.2706 0.6188 3.3087 8.6510 3.3805 0.7856 23.7068 1.0000 0.4110 0.2126 0.1459 0.2305RAG-1 GTR + I + C 0.3280 1.6718 1.2761 4.1305 0.6188 1.4669 5.6222 1.0000 0.3329 0.2123 0.1756 0.2793Tyr HKY + I + C 0.3356 1.1107 Ti/tv ratio 2.5277 0.2560 0.2399 0.2111 0.2931
Table 4Statistical support for three proposed DNA sequence data partitioning schemes for phylogenetic analyses. 1-way: a single model for concatenated 5-gene data set. 5-way:partitioning data by gene, i.e., 12S, 16S, COI, RAG-1 and Tyr. 11-way: each protein-coding gene (COI, RAG-1 and Tyr) partitioned independently by codon position, with ribosomalgenes 12S and 16S as two partitions. Details regarding calculations of Bayes factors and Akaike weights are provided in Section 2.
Partition scheme Free parameters Harmonic mean of log likelihoods Bayes factor Relative Bayes factor AIC Delta AIC Akaike weight
N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx 5
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
Fig. 2. Bayesian consensus tree of the genus Pristimantis based on a 5-gene analysis of 202 frogs. The outgroups Agalychnis callydryas, Litoria caerulea, Diasporus hylaeformis, D.vocator, Craugastor longirostris, Phrynopus auriculatus, P. bracki, Lynchius flavomaculatus, L. nebulanastes, Oreobates cruralis, O. saxatilis are not shown, but their relationships arepresented in Supplementary material Fig. S1. Bootstrap support values are presented for parsimony and likelihood above each branch (upper value corresponds to parsimonyanalysis), with Bayesian posterior probabilities (#100) below each branch. Support values were not presented for any node with low support in all three search strategies. Theareas shaded in gray correspond to the subgenus denoted by the letter H for Hypodyction, Y for Yunganastes and the P. ridens Species Series as defined by Hedges et al. (2008)and P .rubicundus Species Series as defined by Wang et al. (2008). Species groups are indicated by symbols to the right of taxon names, as follows (Hedges et al., 2008): ( )chalceus Species Group, (j) P. conspicillatus Species Group, (.) P. curtipes Species Group, (}) P. devillei Species Group, (s) P. frater Species Group, ( ) P. galdi Species Group, (4)P. lacrimosus Species Group, d P. myersi Species Group, (h) P. orcesi Species Group, ( ) P. orestes Species Group, ( ) P. pardalis Species Group (!) P. peruvianus Species Group,(O) P. surdus Species Group, ( ) P. unistrigatus Species Group. Taxa for which the genus is not indicated belong to the Pristimantis genus.
6 N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx 7
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
America to Central America (Fig. 3). Most of these invasions oc-curred in the Miocene, and according to the 95% CI on divergencetimes obtained under all five calibration schemes applied to theroot age of the tree, at least eight of these invasions into CentralAmerica occurred well before 3.5 Ma (Table 6), i.e., well beforethe canonical date for the completion of the Central American land
bridge (Coates and Obando, 1996). Assuming that current taxon-omy accurately reflects species boundaries in Pristimantis, in situdiversification within Central America was limited to the P. pardalisSpecies Group and the P. cerasinus + P. aff. cruentus clade.
4. Discussion
This study provides five main insights into the history and tax-onomy of the genus Pristimantis in South and Central America.First, we corroborate that the geographic origin of the genus isSouth American. Second, the presence of Pristimantis in CentralAmerica is a result of multiple colonization events from SouthAmerica (Fig. 2), with statistically significant ancestral-area recon-structions suggesting a minimum of 11 independent invasions(Fig. 3), although the data are also compatible with a higher num-ber of invasions. Third, minimum divergence times for the crownage of each independent Central American lineage of Pristimantisshow that at least eight lineages were present in Central Americawell before the hypothesized closure of the Isthmus at 3.5–3.1 Ma (Fig. 3, Table 6), suggesting either multiple over-water dis-persal events between continents or greater subareal connectivitythan previously appreciated. Fourth, much of the divergence with-in Central America was cryptic, with deep splits among lineagesstill identified as conspecific, except for the in situ radiation ofthe 3-species P. pardalis Species Group (Fig. 3). Finally, most ofthe taxonomic groups, series and subgenera currently recognizedare not supported by our molecular phylogenetic analyses.
4.1. Multiple invasions
The present data set includes all 12 named species of Pristiman-tis known from Central America prior to 2010, and confirms that
Fig. 2 (continued)
Table 5Results of tests for monophyly for each taxonomic group within Pristimantis for which>1 species was available (Fig. 2 and Supplementary material Table S2). For each groupa constrained topology was compared to the unconstrained ML topology using theShimodaira–Hasegawa test (Shimodaira and Hasegawa, 1999). Results are presentedby subgenus, Species Series and Species Group. Subgenera or groups with only onesampled representative are designated N/A, since their monophyly could not beevaluated. The sampled species for each group are listed in Supplementary materialTable S2.
8 N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
South America is the ancestral area for Pristimantis (Duellman,2001; Heinicke et al., 2007; Savage, 2002; Vanzolini and Heyer,1985). The ancestral-area analyses revealed a history of multipleindependent invasions by congeners into Central America (Fig. 3),which had not been demonstrated previously. Discrete dispersalevents from South to Central America include eight species nomi-nally shared between continents, i.e., P. achatinus, P. caryophyllac-eus, P. cf. taeniatus 1, P. cruentus, P. gaigei, P. moro, P. ridens and P.taeniatus. In all eight cases the apparent sister lineage is SouthAmerican, thus DEC, and parsimony inferences suggest that eachset of Central American populations represents an independentinvasion without subsequent speciation. Given that the presentdata set includes nearly all of the Pristimantis diversity in CentralAmerica, the number of invasions is unlikely to change with fur-ther sampling of either continent since what is missing from ourcurrent phylogeny is more South American diversity. We may beunderestimating the degree of in situ diversification, however,since our analyses revealed deep conspecific divergences withinCentral American species. If further systematic work were to splitthese wide-ranging taxa, we would have to revise our estimate ofthe importance of in situ speciation events following invasion fromSouth America.
In addition to the eight wide-spread species, three invasions aremanifested by endemic Central American lineages. Pristimantismuseosus is a Panamanian endemic with its sister lineage rangingfrom northern Colombia to Ecuador. The clade containing P. cerasi-nus + P. aff. cruentus is the sister taxon to a clade comprising theSouth American P. viejas and the Central + South American P. aff.taeniatus. The third invasion involved the ancestor of the P. pardalisSpecies Group that subsequently diversified into at least three spe-cies: P. altae, P. pirrensis and P. pardalis (the latter appeared in Fig. 3as a polyphyletic taxon).
The hypothesis of 11 independent invasions within a singlegenus may seem unexpectedly high, yet in the context of the manyspecies of the genus Pristimantis, these events are phylogeneticallywell-separated and supported by both parsimony and DEC infer-ences (Fig. 3). Furthermore, independent colonization events ofNorth and South America by multiple lineages within a taxon areknown from additional groups, such as mammals (Webb and Ran-cy, 1996), birds (Weir et al., 2009), dendrobatid frogs (Santos et al.,2009). The frequency of dispersal events within and among groups
would suggest that colonization was relatively easy and may argueagainst the need to invoke rare ‘‘sweepstakes’’ processes (Simpson,1940; Cody et al., 2010).
The reliability of biogeographic inference depends upon ade-quate lineage sampling and robust phylogenetic inference, as wellas low rates of evolution (Donoghue and Moore, 2003). Given thatwe have sampled extensively the diversity of Pristimantis in Pan-ama, and that the phylogenetic hypothesis presented here hasstrong statistical support for most clades participating in invasionsfrom South to Central America, increased sampling from SouthAmerica would not reduce the minimum number of independentcolonization events inferred here. Adding species from the sourceregion to the phylogeny could actually increase the number of in-ferred dispersal events if, for example, what we inferred to be astrictly Central American clade (e.g., the P. pardalis Species Group)actually contained unsampled South American lineages.
4.2. Temporal framework
Most of the dispersal events (8 of 11) occurred before the gen-erally accepted date for the completion of the Central Americanland bridge (Coates and Obando, 1996; Coates et al., 2004), sug-gesting that prior to 3.5–3.1 Ma whatever oceanic gaps existed inthe otherwise continuous land span did not prevent amphibiansfrom dispersing between continents (Weigt et al., 2005) (Table 6).Fossil mammal data from Central Panama suggest that southernCentral America had a continuous connection with North Americaduring the middle Miocene (Whitmore and Stewart, 1965; Kirbyand MacFadden, 2005). This peninsula might have received earlyanuran colonists from South America, such as P. ridens (Wanget al., 2008) and the túngara frog (Weigt et al., 2005), during thelate Miocene. Pristimantis could have arrived in Central Americabefore the completion of a land bridge by rafting (Vences et al.,2004), or during the end of the Miocene when the sea level wasapproximately 60 m below today’s level (Perdices et al., 2002).
Taking advantage of our phylogeographic sampling of conspe-cific populations within Panama, we were able to infer the mini-mum ages of colonization for these taxa. Using the minimum ofthe 95% credibility interval for the TMRCA of conspecific CentralAmerican samples, i.e., crown ages (Fig. 3), we observed that theMRCA for 8 of 11 species and clades was likely already present
Table 6Estimated crown ages in millions of years ago (Ma) for taxa that contain Central American representatives obtained from five alternative calibration intervals for the age of theroot node of the molecular phylogeny (Fig. 3). Ages were estimated from Bayesian relaxed clock analyses implemented in the software BEAST (see Section 2 for details). BecausePristimantis museosus was represented in our data set by a single individual, we report the credibility interval for its crown age as zero to the mean posterior estimate of theTMRCA of its sister lineage. Asterisks indicate divergence time estimates that would be compatible with a Central American species having colonized Central American after theclosure of the Isthmus of Panama, assuming the canonical date of 3.5–3.1 Ma for this geological event. Thus, at least 8 of 11 clades entered Central America prior to 4 Ma,regardless of the hypothesized root age.
Taxon or clade inCentral America
Estimated age (Ma) Bayesian 95% credibility interval (Ma)
N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx 9
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
Fig. 3. A chronogram of Pristimantis derived from a relaxed-clock Bayesian analysis using the software BEAST and assuming a mean root age of 57 million years ago (Ma) witha standard deviation of 9 million years (alternative calibration results are presented in Table 6). Scale along the bottom indicates time in Ma. Branch colors reflectbiogeographic designations (for species at tips) and ancestral state estimates (for internal nodes), estimated under the DEC model of Ree and Smith (2008). Red is CentralAmerican, black is South American and blue is widely distributed (i.e., found in both regions). For tip branches, state changes are shown (arbitrarily) at the mid-point of eachbranch. Dashed lines indicate uncertain reconstruction of ancestral state while solid lines indicate that all alternative reconstructions fell >2 log-likelihood units lower thanthe MLE (Ree and Smith, 2008), typically much lower. Tip labels are same as in Fig. 2. The green vertical line indicates the hypothesized 3.5–3.1 Ma completion of the Isthmusof Panama. Gray horizontal bars indicate 95% credibility intervals for the divergence time of the genus Pristimantis and the 11 lineages with representation in Central America.Asterisks on nodes indicate estimated posterior probabilitiesP0.95 for the presence of the corresponding clade according to BEAST. Thin branches on tree lead to samples notused in DEC analyses due to limitations of the software.
10 N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
on the Panamanian Isthmus prior to 4 Ma (Table 6). Although theestimated ages of colonization of Pristimantis species presentedhere seem quite old (7.6–13 Ma), they agree with dates estimatedfor other amphibians such as túngara frogs at 6–10 Ma (Weigtet al., 2005), treefrogs at 3–20 Ma (Moen et al., 2009), dendrobatidfrogs during the Miocene 5–23 Ma (Santos et al., 2009), and sala-manders of the predominantly South American Bolitoglossa adsper-
sa group at 11–18 Ma (Wiens et al., 2007), as well as othervertebrate groups such as primary freshwater fishes at 4–7 Ma(Bermingham and Martin, 1998) and vipers at 0.8–22.8 (Castoeet al., 2009; Daza et al., 2010; Zamudio and Greene, 1997).
Using molecular data to date events has many possible sourcesof uncertainty, including rate variation among lineages, accuracy ofcalibration points, saturation of nucleotide positions, and genetic
Fig. 3 (continued)
N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx 11
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
polymorphism in ancestral populations (Arbogast et al., 2002;Rutschmann, 2006). To account for these possible sources oferror we employed a relaxed-clock method based on threeindependent gene regions (mitochondrial DNA plus twopresumably unlinked nuclear genes). Uncertainty in diver-gence-time estimates due to rate variation and mutational sto-chasticity is described by the 95% confidence intervals aroundeach node. The most important assumption behind our diver-gence time analysis, therefore, is our calibration intervals. Werepeated the temporal analyses assuming five publishedhypotheses for the root age of our tree. For each hypothesis,we also applied wide intervals to account for the uncertaintyin each of these secondary calibrations (Ho, 2007). Our resultswere robust to assumptions of root age, as all five analyses sup-ported the early colonization of Central America at least 8 of 11lineages of Pristimantis (Table 6).
The growing list of terrestrial animal lineages thought to havecolonized Central from South America well before 3.5 Ma pre-sents somewhat of a conundrum. Studies such as the presentone suggest a high degree of connectivity between continentsas long ago as the Miocene, and various geological scenarios con-firm the presence of well-developed yet dis-connected islands inplace at this time (Van Andel et al., 1971; Iturralde and Mac-Phee, 1999; Coates et al., 2004). More recent geological studiesargue for even earlier dates for the initiation of the formationof the Isthmus (Farris et al., 2011). Data from marine fossils,however, argue against the presence of a complete Miocene landconnection separating the present-day Caribbean and Pacificoceans (e.g., O’Dea et al., 2007). Assuming that the variousamphibian species that entered Central America during the Mio-cene did not cross open ocean, how could we posit terrestrialdispersal without invoking a continuous land bridge dividingthe ocean? One possible resolution of this problem would beto acknowledge the highly dynamic nature of the formation ofthe Isthmian landbridge, including changes in see level and pos-sibly in the height of geological formation, e.g., cooling and sink-ing (Coates et al., 2003). The islands comprising the nascentlandbridge may have connected and dis-connected over timesuch that terrestrial organisms could have moved between con-tinents even before the final completion of the Isthmus.
4.3. Formation of Central American frog communities
The Central American community of Pristimantis was formedby a mixture of colonization and speciation, with the former pre-dominating. The Central American endemic P. pardalis SpeciesGroup (P. altae, P. pardalis and P. pirrensis; Wang et al., 2008)provides a clear case of speciation following colonization(Fig. 3), having diverged from its nearest (according to our sam-pling) South American relatives an estimated 17 Ma (12–22 Ma).The second case of diversification in situ is formed by P. aff.cruentus (FMNH 257553) and P. cerasinus, having diverged fromits nearest South and Central American relatives an estimated12 (8–18 Ma). The third endemic lineage, P. museosus, dispersedinto Central America 0–14 Ma, followed by an extinction eventof the corresponding source population in South America (Reeet al., 2005). The remaining eight Central American species ofPristimantis in our analysis (including P. aff. taeniatus 1) haveconspecific populations in Colombia, South America. Their pres-ence in Central America represents at least eight dispersal eventsapparently without subsequent speciation, though this impres-sion could be due to an incomplete taxonomy of Central Amer-ican Pristimantis (Crawford et al., 2010a).
Our assessment of the relative contributions of colonizationversus in situ speciation in the formation of the Central AmericanPristimantis community is not affected by the accuracy of the geo-
logical model assumed for the formation of the Isthmus (Coatesand Obando, 1996; Kirby and MacFadden, 2005). Rather, the geo-logical model suggests whether colonization events required cross-ing open ocean. For six of the eight widespread species sharedbetween South and Central America, the minimum 95% credibleinterval for the crown age of just the Central American populationsis >4 million years (Table 6, Fig. 3). Our analyses suggest, therefore,that most of the Central American Pristimantis fauna was in placeprior to the canonical 3.5–3.1 Ma date for the closure of the Isth-mus. The only widespread species that potentially entered subse-quent to this date would have been P. gaigae (reaching CostaRica) and P. achatinus (found only in eastern-most Panama).Regardless of precisely how or when it formed, the Isthmus of Pan-ama has had a rich biotic history that has fostered numerous ende-mic lineages (Bermingham and Martin, 1998; Crawford et al.,2010b; Ibáñez and Crawford, 2004; Reeves and Bermingham,2006; Wang et al., 2008).
4.4. Taxonomic implications
Among the two subgenera and 12 taxonomic series and groupswithin Pristimantis evaluated here, we found support for threegroups: the P.myersi Species Group sensu Hedges et al. (2008) (Ta-ble 5, Supplementary material Table S2), the P. pardalis SpeciesGroup sensuWang et al. (2008) and the P. rubicundus Species Seriessensu Crawford et al. (2010b) (Table 5). The subgeneric taxonomyof Pristimantis is clearly flawed. The limited taxon sampling (134of 437 species) makes it difficult to revise the taxonomy of Pristi-mantis, however, and we refrain from re-defining groups untilDNA and other relevant data become available for a larger propor-tion of the genus Pristimantis.
5. Conclusions
Our dense phylogenetic sampling and likelihood-based bio-geographic analysis of the genus Pristimantis reveals that CentralAmerica was colonized through multiple invasions, most ofwhich occurred 6–12 Ma (Table 6). The similarity of dates ob-tained in this study, and the fact that these dates match thoseof previous studies based on independent lineages and indepen-dent assumptions (e.g., Weigt et al., 2005), suggests that the tra-ditional geological scenario for the formation of the Isthmus ofPanama or its presumed impact on terrestrial biogeographymay have to be reconsidered. The diversity of Pristimantis inCentral America would seem to have been driven more by colo-nization than by in situ diversification, though the large intraspe-cific divergences suggest that the current taxonomy ofPristimantis may be far from complete.
Acknowledgments
We are grateful to the following individuals and institutionswho provided specimens, permits, and tissues necessary for thisstudy: C. Jaramillo at the Círculo Herpetológico de Panamá, M.P.Ramirez of the Colección Herpetológica, Universidad Industrial deSantander in Colombia, E. Muñoz and V. Páez at the Museo de Her-petología de la Universidad de Antioquia in Colombia, J.M. Hoyos atthe Museo Javeriano de Historia Natural de la Pontificia Universi-dad Javeriana in Bogotá, Colombia, Karen Siu-Ting and César Agu-ilar of the Museo de Historia Natural de la Universidad NacionalMayor de San Marcos en Lima, Peru, K.R. Lips at the University ofMaryland, USA, E.N. Smith of the Amphibian and Reptile DiversityResearch Center at the University of Texas at Arlington, USA, and L.Coloma formerly of the Museo de Zoología de la Pontificia Univers-idad Católica del Ecuador.
12 N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
Research permits in Colombia were issued by the Ministeriodel Medio Ambiente (No. 13 del 21 de diciembre 2006 toN.R.P.-S and No. 11 del 18 de diciembre de 2006 to A.J.C.). Thisstudy is included in the ‘‘Contrato de Acceso a Recursos Genéti-cos No. 0040, del 11 de enero de 2008’’ to N.R.P.-S. and Resolu-ción Número 1816 del 22 de septiembre 2009 (contrato No. 31)to A. Amézquita granted by the Ministerio del Medio AmbienteVivienda y Desarrollo Territorial, Colombia. Field help was kindlyprovided by J.M. Renjifo. We are indebted to J.D. Lynch for helpin identification of specimens. Collections in Panama were con-ducted under permit Nos. SE/A-083-2001 and SE/A-37-07 toR.I.D. and SE/AP-7-07 to E.B. by the Autoridad Nacional delAmbiente, with field assistance generously provided by C. Jara-millo, E. Griffith, D. Medina, R. Brenes, J. de Alba, S. Flechas,Ach. Batista, S. Lanckowsky, D. Reznick, J.J. Wiens, R. Puschendorfand G. Berguido of the Reserva Natural Privada Chucantí. Collec-tions in Costa Rica were made possible by Ministerio del Ambi-ente y Energía permit Nos. 024-2002-OFAU and 163-2003-OFAUto A.J.C. and with the assistance of F. Bolaños, G. Chaves, R. Ro-jas, R. Puschendorf and B. Kubicki of the Costa Rican AmphibianResearch Center (CRARC).
We are indebted to D.S. Moen and C. Sanín for help with ances-tral areas reconstruction analysis, G. Grajales and M. González forhelp in the laboratory, the Biom|ics Lab especially A. Paz for assis-tance and feedback, and to C. Sarmiento and M. Escobar for helpwith graphics.
This work was supported by grants from Colombia’s InstitutoFrancisco José de Caldas for the Advancement of Science and Tech-nology (Colciencias; Convenio No. 074 from 2006), an Adelante Fel-lowship from the Smithsonian Tropical Research Institute, and theResearch Committee of the Faculty of Sciences at Universidad delos Andes. Valuable comments on the manuscript were providedby J. Daza, A. Larson and one anonymous reviewer. This researchformed the basis of the masters’ thesis of N.R.P.-S. presented tothe Departamento de Ciencias Biológicas at the Universidad delos Andes.
Appendix A
All species are members of the genus Pristimantis, except forthe outgroups: Craugastor and Diasporus. An asterisk ! in the spe-cies column indicates that this sample was not included in theDEC analysis. Ten specimens are still awaiting accession intoinstitutional natural history collections. For each specimen, mu-seum voucher, source, locality and GenBank accession numberare reported. Acronyms for museums are: ANDES = Museo de His-toria Natural ANDES, Colombia; CH = Circulo Herpetológico de Pa-namá, Panama; FMNH = Field Museum Natural History, USA;MHUA = Museo de Herpetología de la Universidad de Antioquia,Colombia, MUSM = Museo de Historia Natural de la UniversidadNacional Mayor de San Marcos en Lima, Peru; MVUP = Museode Vertebrados de la Universidad de Panamá, Panama; QCAZ = -Museo de Zoología de la Pontificia Universidad Católica del Ecua-dor, Ecuador; SIUC-H = Southern Illinois University at Carbondale,USA; UCR = Universidad de Costa Rica, Costa Rica; UTA-A = Uni-versity of Texas at Arlington, USA; UIS-H = Colección Herpetológ-ica, Universidad Industrial de Santander, Colombia. Theabbreviations for the individuals’ field series are as follows: AJ-C = Andrew J. Crawford; CJD = Claudia Juliana Dulcey; EMM = Eli-ana Maria Muñoz; ENS = Eric N. Smith; KRL = Karen R. Lips;KST = Karen Siu-Ting; MBH = Michael B. Harvey; NRPS = Nelsy Ro-cio Pinto-Sánchez; PDG = Paul David Gutiérrez; RC = Rances Caice-do. The collection locality abbreviations are: EB = EstaciónBiológica; PN = Parque Nacional; RB = Reserva Biológica. N/A = data for corresponding gene not available for that sample.
Species
Institutiona
lvo
uche
rnu
mbe
r
Field
colle
ction
numbe
r
Coun
try
Dep
artm
ent/
prov
ince
Mun
icipality/
locality
Latitude
Long
itud
eMitoc
hond
rial
gene
sNuc
lear
gene
s
COI
16S
12S
Rag1
Tyr
acha
tinu
sMVUP18
59AJC
0573
Pana
ma
Darién
Cana
maincamp
8.05
"77
.58
JN99
1349
JN99
1420
JN99
1485
JQ02
5168
JN99
1552
aff.altamazon
icus
MUSM
2691
1AJC
2005
Peru
Huá
nuco
Pang
uana
"9.6
"74
.94
JN99
1350
JN99
1421
N/A
N/A
N/A
aff.altamazon
icus
⁄MUSM
2691
2AJC
2006
Peru
Huá
nuco
Pang
uana
"9.6
"74
.94
JN99
1351
JN99
1422
N/A
JQ02
5169
N/A
aff.crue
ntus
FMNH
2575
53AJC
0217
Pana
ma
Chiriquí
Fortun
a8.75
"82
.22
JN99
1352
JN99
1423
JN99
1486
JQ02
5170
JN99
1553
aff.taen
iatus1
ANDES
-A63
5AJC
1191
ColombiaCh
ocó
Nuq
uí10
.22
"84
.6JN
9913
54JN
9914
25JN
9914
88N/A
N/A
aff.taen
iatus3
ANDES
-A63
8AJC
1353
ColombiaTo
lima
Falán
5.12
"74
.95
JN99
1356
N/A
JN99
1491
JQ02
5173
JN99
1556
aff.taen
iatus3⁄
ANDES
-A64
0AJC
1368
ColombiaTo
lima
Falán
5.12
"74
.95
JN99
1357
JN99
1428
JN99
1492
N/A
JN99
1557
aff.taen
iatus1
Not
catalogu
edAJC
1683
Pana
ma
Darién
Cana
maincamp
7.93
"77
.72
JN99
1355
JN99
1426
JN99
1489
N/A
JN99
1555
aff.taen
iatus3
ANDES
-ANRP
SCo
lombiaAntioqu
iaAno
rí6.98
"75
.13
JN99
1359
JN99
1430
JN99
1494
JQ02
5172
JN99
1559
(con
tinu
edon
next
page)
N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx 13
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
Supplementary data associated with this article can be found inthe online version, at doi:10.1016/j.ympev.2011.11.022.
References
Akaike, H., 1973. Information theory and an extension of the maximum likelihoodprinciple. In: Petrov, B.N., Csaki, F. (Eds.), 2nd International Symposium onInformation Theory. Akademiai Kiado, Budapest, Hungary, pp. 267–281.
Amos, W., Whitehead, H., Ferrari, M., Glockner Ferrari, D., Payne, R., Gordon, J., 1992.Restrictable DNA from sloughed cetacean skin; its potential for use inpopulation analysis. Mar. Mamm. Sci. 8, 275–283.
AmphibiaWeb, 2011. Information on Amphibian Biology and Conservation.Berkeley, California. <http://amphibiaweb.org/> (web application, accessed13.12.11).
Arbogast, B.S., Edwards, S.V., Wakeley, J., Beerli, P., Slowinski, J.B., 2002. Estimatingdivergence times from molecular data on phylogenetic and population genetictimescales. Ann. Rev. Ecol. Syst. 33, 707–740.
Barker, F.K., Lutzoni, F.M., 2002. The utility of the incongruence length differencetest. Syst. Biol. 51, 625–637.
Bermingham, E., Martin, A.P., 1998. Comparative mtDNA phylogeography ofneotropical freshwater fishes: testing shared history to infer the evolutionarylandscape of lower Central America. Mol. Ecol. 7, 499–517.
Bossuyt, F., Milinkovitch, M.C., 2001. Amphibians as indicators of early Tertiary‘‘Out-of-India’’ dispersal of vertebrates. Science 292, 93–95.
Brandley, M., Schmitz, A., Reeder, T., 2005. Partitioned Bayesian analyses, partitionchoice, and the phylogenetic relationships of scincid lizards. Syst. Biol. 54, 373–390.
Castoe, T., Parkinson, C., 2006. Bayesian mixed models and the phylogeny ofpitvipers (Viperidae: Serpentes). Mol. Phylogenet. Evol. 39, 91–110.
Castoe, T., Sasa, M., Parkinson, C., 2005. Modeling nucleotide evolution at themesoscale: the phylogeny of the Neotropical pitvipers of the Porthidium group(Viperidae: Crotalinae). Mol. Phylogenet. Evol. 37, 881–898.
Castoe, T.A., Daza, J.M., Smith, E.N., Sasa, M.M., Kuch, U., Campbell, J.A., Chippindale,P.T., Parkinson, C.L., 2009. Comparative phylogeography of pitvipers suggests aconsensus of ancient Middle American highland biogeography. J. Biogeogr. 36,88–103.
Castresana, J., 2000. Selection of conserved blocks from multiple alignments fortheir use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552.
Clark, J.R., Ree, R.H., Alfaro, M.E., King, M.G., Wagner, W.L., Roalson, E.H., 2008. Acomparative study in ancestral range reconstruction methods: retracing theuncertain histories of insular lineages. Syst. Biol. 57, 693–707.
Coates, A.G., Obando, J.A., 1996. The geologic evolution of the Central AmericanIsthmus. In: Jackson, J.B.C., Budd, A.F., Coates, A.G. (Eds.), Evolution andEnvironment in Tropical America. University of Chicago Press, Chicago, pp.21–56.
Coates, A.G., Aubry, M.-P., Berggren, W.A., Collins, L.S., Kunk, M., 2003. EarlyNeogene history of the Central American arc from Bocas del Toro, westernPanama. GSA Bullet. 115, 271–287.
Coates, A.G., Collins, L.S., Aubry, M.-P., Berggren, W.A., 2004. The geology of theDarien, Panama, and the late Miocene–Pliocene collision of the Panama arc withnorthwestern South America. GSA Bullet. 116, 1327–1344.
Crawford, A.J., Smith, E.N., 2005. Cenozoic biogeography and evolution in direct-developing frogs of Central America (Leptodactylidae: Eleutherodactylus) asinferred from a phylogenetic analysis of nuclear and mitochondrial genes. Mol.Phylogenet. Evol. 35, 536–555.
Crawford, A.J., Bermingham, E., Polanía, C., 2007. The role of tropical dry forest as along-term barrier to dispersal: a comparative phylogeographical analysis of dryforest tolerant and intolerant frogs. Mol. Ecol. 16, 4789–4807.
Crawford, A.J., Lips, K.R., Bermingham, E., 2010a. Epidemic disease decimatesamphibian abundance, species diversity, and evolutionary history in thehighlands of central Panama. Proc. Natl. Acad. Sci. USA 107, 13777–13782.
Crawford, A.J., Ryan, M.J., Jaramillo, C.A., 2010b. A new species of Pristimantis(Anura: Strabomantidae) from the Pacific coast of the Darien Province, Panama,with a molecular analysis of its phylogenetic position. Herpetologica 66, 192–206.
Dacosta, J.M., Klicka, J., 2008. The Great American Interchange in birds: aphylogenetic perspective with the genus Trogon. Mol. Ecol. 17, 1328–1343.
Daza, J.M., Castoe, T.A., Parkinson, C.L., 2010. Using regional comparativephylogeographic data from snake lineages to infer historical processes inMiddle America. Ecography 33, 343–354.
N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx 17
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
Folmer, O., Black, M., Hoeh, W., Lutz, R., Vrijenhoek, R., 1994. DNA primers foramplification of mitochondrial cytochrome c oxidase subunit I from diversemetazoan invertebrates. Mol. Mar. Biol. Biotechnol. 3, 294–299.
Frost, D., 2009. Amphibian Species of the World: An Online reference. Version 5.3.<http://research.amnh.org/herpetology/amphibian/index.html>.
Frost, D.R., Grant, T., Faivovich, J., Baina, R.H., Haas, A., Haddad, C.F., de Sá, R.O.,Channing, A., Wilkinson, M., Donnellan, S.C., Raxworthy, C.J., Campbell, J.A.,Blotto, B.L., Moler, P., Drewes, R.C., Nussbaum, R.A., Lynch, J.D., Green, D.M.,Wheeler, W.C., 2006. The amphibian tree of life. Bull. Am. Mus. Nat. Hist. 297, 1–291.
Guayasamin, J.M., Castroviejo-Fisher, S., Ayarzagüena, J., Trueb, L., Vilà, C., 2008.Phylogenetic relationships of glassfrogs (Centrolenidae) based on mitochondrialand nuclear genes. Mol. Phylogenet. Evol. 48, 574–595.
Hanken, J., Wake, D., 1982. Genetic differentiation among plethodontid salamanders(genus Bolitoglossa) in Central and South America: implications for the SouthAmerican invasion. Herpetologica 38, 272–287.
Hasegawa, M., Kishino, H., Yano, T.-a., 1985. Dating of the human-ape splitting by amolecular clock of mitochondrial DNA. J. Mol. Evol. 22, 160–174.
Hedges, S.B., Duellman, W.E., Heinicke, M.P., 2008. New World direct-developingfrogs (Anura: Terrarana): molecular phylogeny, classification, biogeography,and conservation. Zootaxa 1737, 1–182.
Heinicke, M., Duellman, W., Trueb, L., Means, D., MacCulloch, R., Hedges, S., 2009. Anew frog family (Anura: Terrarana) from South America and an expandeddirect-developing clade revealed by molecular phylogeny. Zootaxa 2211, 1–35.
Heinicke, M.P., Duellman, W.E., Hedges, S.B., 2007. Major Caribbean and CentralAmerican frog faunas originated by ancient oceanic dispersal. Proc. Natl. Acad.Sci. USA 104, 10092–10097.
Ho, S.Y.W., 2007. Calibrating molecular estimates of substitution rates anddivergence times in birds. J. Avian Biol. 38, 409–414.
Ibáñez, R., Crawford, A.J., 2004. A new species of Eleutherodactylus (Anura:Leptodactylidae) from the Darién Province. Panama. J. Herpetol. 38, 240–244.
Iturralde, M.A., MacPhee, R.D.E., 1999. Paleogeography of the Caribbean region:implications for Cenozoic biogeography. Bull. Am. Mus. Nat. Hist. 238, 1–96.
Kass, R.E., Raftery, A.E., 1995. Bayes factors. J. Am. Stat. Assoc. 90, 773–795.Katoh, K., Toh, H., 2010. MAFFT version 6: parallelization of the MAFFT multiple
sequence alignment program. Bioinformatics 26, 1899–1900.Kirby, M., MacFadden, B., 2005. Was southern Central America an archipelago or a
peninsula in the middle Miocene? A test using land-mammal body size.Palaeogeogr. Palaeoclimatol. Palaeoecol. 228, 193–202.
Koepfli, K.P., Gompper, M.E., Eizirik, E., Ho, C.C., Linden, L., Maldonado, J.E., Wayne,R.K., 2007. Phylogeny of the Procyonidae (Mammalia: Carnivora): molecules,morphology and the great American interchange. Mol. Phylogenet. Evol. 43,1076–1095.
Lynch, J.D., Duellman, W.E., 1997. Frogs of the Genus Eleutherodactylus in WesternEcuador. University of Kansas Special Publication 23, pp. 1–236.
Maddison, W.P., Maddison, D.R., 2009. Mesquite 2.74: A Modular System forEvolutionary Analysis <http://mesquiteproject.org>.
Marshall, L.G., Webb, S., Sepkoski Jr., J., Raup, D., 1982. Mammalian evolution andthe great American interchange. Science 215, 1351–1357.
Marshall, L.G., 1988. Land mammals and the Great American Interchange. Am. Sci.76, 380–388.
Maxson, L.R., Myers, C.W., 1985. Albumin evolution in tropical poison frogs(Dendrobatidae): a preliminary report. Biotropica 17, 50–56.
Meyer, C., Paulay, G., 2005. DNA barcoding: error rates based on comprehensivesampling. PLoS Biol. 3, 2229.
Meyer, C.P., Geller, J.B., Paulay, G., 2005. Fine scale endemism on coral reefs:archipelagic differentiation in turbinid gastropods. Evolution 59, 113–125.
Miller, M., Holder, M., Vos, R., Midford, P., Liebowitz, T., Chan, L., Hoover, P.,Warnow, T., 2010. The CIPRES Portals. CIPRES. <http://www.phylo.org/sub_sections/portal (accessed 06.01.10).
Moen, D.S., Smith, S.A., Wiens, J.J., 2009. Community assembly through evolutionarydiversification and dispersal in Middle American treefrogs. Evolution 63, 3228–3247.
Morgan, G., 2005. The great American Biotic Interchange in Florida. Bull. Fla. Mus.Nat. Hist. 45, 271–311.
Nylander, J.A.A., 2004. MrModeltest 2.3. Computer Program Distributed by theAuthor.
O’Dea, A., Jackson, J.B.C., Fortunato, H., Smith, J.T., D’Croz, L., Johnson, K.G., Todd, J.A.,2007. Environmental change preceded Caribbean extinction by 2 million years.Proc. Natl. Acad. Sci. USA 104, 5501–5506.
Padial, J., Castroviejo-Fisher, S., Köhler, J., Domic, E., De la Riva, I., 2007. Systematicsof the Eleutherodactylus fraudator species group (Anura: Brachycephalidae).Herpetol. Monogr. 21, 213–240.
Padial, J.M., de la Riva, I., 2009. Integrative taxonomy reveals cryptic Amazonianspecies of Pristimantis (Anura: Strabomantidae). Zool. J. Linn. Soc. 155, 97–122.
Palumbi, S., Martin, A., Romano, S., McMillan, W., Stice, L., Grabowski, G., 1991. TheSimple Fool’s Guide to PCR, Version 2.0, Privately Published, Univ. Hawaii.
Perdices, A., Bermingham, E., Montilla, A., Doadrio, I., 2002. Evolutionary history ofthe genus Rhamdia (Teleostei: Pimelodidae) in Central America. Mol.Phylogenet. Evol. 25, 172–189.
Posada, D., Buckley, T., 2004. Model selection and model averaging inphylogenetics: advantages of Akaike information criterion and Bayesianapproaches over likelihood ratio tests. Syst. Biol. 53, 793–808.
Posada, D., Crandall, K., 1998. MODELTEST: testing the model of DNA substitution.Bioinformatics 14, 817–818.
Pyron, R.A., Wiens, J.J., 2011. A large-scale phylogeny of Amphibia including over2800 species, and a revised classification of extant frogs, salamanders, andcaecilians. Mol. Phylogen. Evol. (online early).
Rambaut, A., Drummond, A., 2004. Tracer 1.3. University of Edinburgh, Edinburgh,UK <http://tree.bio.ed.ac.uk/software/tracer>.
Ratnasingham, S., Hebert, P.D.N., 2007. BoLD: The barcode of life data system(http://www.barcodinglife.org). Mol. Ecol. Notes 7, 355–364.
Ree, R., Moore, B., Webb, C., Donoghue, M., Crandall, K., 2005. A likelihoodframework for inferring the evolution of geographic range on phylogenetictrees. Evolution 59, 2299–2311.
Ree, R., Smith, S., 2008. Maximum likelihood inference of geographic rangeevolution by dispersal, local extinction, and cladogenesis. Syst. Biol. 57, 4–14.
Reeves, R.G., Bermingham, E., 2006. Colonization, population expansion, and lineageturnover: phylogeography of Mesoamerican characiform fish. Biol. J. Linn. Soc.Lond. 88, 235–255.
Roelants, K., Gower, D.J., Wilkinson, M., Loader, S.P., Biju, S.D., Guillaume, K., Moriau,L., Bossuyt, F., 2007. Global patterns of diversification in the history of modernamphibians. Proc. Natl. Acad. Sci. USA 104, 887–892.
Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes version 3.1: Bayesian phylogeneticinference under mixed models. Bioinformatics 19, 1572–1574.
Rutschmann, F., 2006. Molecular dating of phylogenetic trees: a brief review ofcurrent methods that estimate divergence times. Divers. Distrib. 12, 35–48.
Santos, J.C., Coloma, L.A., Summers, K., Caldwell, J.P., Ree, R., Cannatella, D.C., 2009.Amazonian amphibian diversity is primarily derived from Late Miocene Andeanlineages. PLoS Biol. 7, 448–461.
Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method forreconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425.
Savage, J.M., 1982. The enigma of the Central American herpetofauna: dispersal orvicariance? Ann. Mo. Bot. Gard. 69, 464–547.
Savage, J.M., 2002. The Amphibians and Reptiles of Costa Rica: A Herpetofaunabetween Two Continents, Between Two Seas. The University of Chicago Press,Chicago, IL.
Shimodaira, H., Hasegawa, M., 1999. Multiple comparisons of log-likelihoods withapplications to phylogenetic inference. Mol. Biol. Evol. 16, 1114–1116.
Simpson, G.G., 1940. Mammals and land bridges. J. Washington Acad. Sci. 30, 137–163.
Simpson, G.G., 1980. Splendid Isolation: The Curious History of South AmericanMammals, Yale University Press, New Haven.
Smith, M.A., Poyarkov Jr., N.A., Hebert, P.D.N., 2008. CO1 DNA barcodingamphibians: take the chance, meet the challenge. Mol. Ecol. Resources 8,235–246.
Stamatakis, A., Hoover, P., Rougemont, J., 2008. A rapid bootstrap algorithm for theRAxML web servers. Syst. Biol. 57, 758–771.
Swofford, D., 2002. PAUP! 4.0b10. Phylogenetic Analysis Using Parsimony (! andOther Methods). Sinauer Associates, Sunderland, Massachusetts.
Van Andel, T.H., Heath, G.R., Malfait, B.T., Hendricks, D.F., Ewing, J.I., 1971.Tectonics of the Panama Basin, eastern Pacific. J. Washington Acad. Sci. 82,1489–1508.
Vanzolini, P.E., Heyer, W.R., 1985. The American herpetofauna and the interchange.In: Stehli, F.G., Webb, S.D. (Eds.), The Great American Biotic Interchange.Plenum Press, New York, pp. 475–487.
Vences, M., Kosuch, J., Rödel, M., Lötters, S., Channing, A., Glaw, F., Böhme, W., 2004.Phylogeography of Ptychadena mascareniensis suggests transoceanic dispersal ina widespread African Malagasy frog lineage. J. Biogeogr. 31, 593–601.
Wang, I.J., Crawford, A.J., Bermingham, E., 2008. Phylogeography of the Pygmy RainFrog (Pristimantis ridens) across the lowland wet forests of isthmian CentralAmerica. Mol. Phylogenet. Evol. 47, 992–1004.
Webb, S., 1978. A history of savanna vertebrates in the New World. Part II: SouthAmerica and the Great Interchange. Annu. Rev. Ecol. Syst. 9, 393–426.
Webb, S.D., Rancy, A., 1996. Late Cenozoic evolution of the Neotropical mammalfauna. In: Jackson, J.B.C., Budd, A.F., Coates, A.G. (Eds.), Evolution andEnvironment in Tropical America. University of Chicago, Chicago, pp. 335–358.
Weigt, L.A., Crawford, A.J., Rand, A.S., Ryan, M.J., 2005. Biogeography of the túngarafrog, Physalaemus pustulosus: a molecular perspective. Mol. Ecol. 14, 3857–3876.
Weir, J.T., Bermingham, E., Schluter, D., 2009. The Great American Biotic Interchangein birds. Proc. Natl. Acad. Sci. USA 106, 21737–21742.
Whitmore Jr., F.C., Stewart, R.H., 1965. Miocene mammals and Central Americanseaways. Science 148, 180–185.
Wiens, J., 1998. Combining data sets with different phylogenetic histories. Syst. Biol.47, 568–581.
Wiens, J., 2007. Global patterns of diversification and species richness inAmphibians. Am. Nat. 170, S86–S106.
Wiens, J., 2011. Re-evolution of lost mandibular teeth in frogs after more than 200million years, and re-evaluating Dollo’s law. Evolution 65, 1283–1296.
18 N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022
Wiens, J., Parra-Olea, G., García-París, M., Wake, D.B., 2007. Phylogenetic historyunderlies elevational biodiversity patterns in tropical salamanders. Proc. Roy.Soc. B 274, 919–928.
Wiens, J., Pyron, R.A., Moen, D.S., 2011. Phylogenetic origins of local-scalediversity patterns and the causes of Amazonian megadiversity. Ecol. Lett. 14,643–652.
Zamudio, K., Greene, H., 1997. Phylogeography of the bushmaster (Lachesis muta:Viperidae): implications for neotropical biogeography, systematics, andconservation. Biol. J. Linn. Soc. Lond. 62, 421–442.
Zeh, J.A., Zeh, D.W., Bonilla, M.M., 2003. Phylogeography of the harlequin beetle-riding pseudoscorpion and the rise of the Isthmus of Panama. Mol. Ecol. 12,2759–2769.
N.R. Pinto-Sánchez et al. /Molecular Phylogenetics and Evolution xxx (2011) xxx–xxx 19
Please cite this article in press as: Pinto-Sánchez, N.R., et al. The Great American Biotic Interchange in frogs: Multiple and early colonization of CentralAmerica by the South American genus Pristimantis (Anura: Craugastoridae). Mol. Phylogenet. Evol. (2011), doi:10.1016/j.ympev.2011.11.022