Seri Indian traditional knowledge and molecular biology agree: no express train for island-hopping spiny-tailed iguanas in the Sea of Cortés

Transcription

1 Journal of Biogeography (J. Biogeogr.) (2011) 38, ORIGINAL ARTICLE Seri Indian traditional knowledge and molecular biology agree: no express train for island-hopping spiny-tailed iguanas in the Sea of Cortés Christina M. Davy 1,2, Fausto R. Méndez de la Cruz 3, Amy Lathrop 2 and Robert W. Murphy 1,2,4 1 Department of Ecology and Evolutionary Biology, University of Toronto, 25 Wilcocks Street, Toronto, ON M5S 3B2, Canada, 2 Department of Natural History, Royal Ontario Museum, 100 Queen s Park, Toronto, ON M5S 2C6, Canada, 3 Laboratorio de Herpetología, Instituto de Biología, Universidad Nacional Autónoma de México, AP , CP 04510, México, DF, Mexico, 4 State Key Laboratory of Genetic Resources and Evolution, Kunming Institute of Zoology, The Chinese Academy of Sciences, Kunming , China ABSTRACT Aim The role of human activities in species biogeography can be difficult to identify, but in some cases molecular techniques can be used to test hypotheses of human-mediated dispersal. A currently accepted hypothesis states that humans mediated the divergence of two species of spiny-tailed iguanas in the Ctenosaura hemilopha species complex, namely C. conspicuosa and C. nolascensis, which occupy islands in the Sea of Cortés between the peninsula of Baja California and mainland Mexico. We test an alternative hypothesis that follows the traditional knowledge of the Seri Indians and states that the divergence of these species was not mediated by humans. Location Mexico, including Baja California, Sonoran and Sinaloan coastal regions, and Isla San Esteban and Isla San Pedro Nolasco in the Sea of Cortés. Methods We analysed mitochondrial (cytochrome b and cytochrome c oxidase subunit III) DNA sequences from four species in the C. hemilopha species complex. Maximum parsimony and Bayesian inference were used to infer matriarchal genealogical relationships between the species and several outgroup taxa. Bayesian methods were used to estimate divergence times for the major nodes on the trees based on previously published, fossil-calibrated priors. Results Our analysis indicated that lineages within the C. hemilopha species complex diverged long before human colonization of the Americas. The divergence of C. nolascensis and C. conspicuosa could not be attributed to Seri translocations. The matriarchal genealogy of the species complex currently defies a simple biogeographical interpretation. Main conclusions We conclude that humans did not mediate the divergence of C. nolascensis and C. conspicuosa. This conclusion is consistent with the traditional knowledge of the Seri people. These results demonstrate the utility of molecular techniques in investigating potential cases of human-mediated dispersal of plants and animals, and reinforce the importance of considering traditional knowledge in the formation of scientific hypotheses and the interpretation of results. Correspondence: Christina M. Davy, c/o Department of Natural History, Royal Ontario Museum, 100 Queens Park, Toronto, ON M5S 2C6, Canada. christina.davy@utoronto.ca Keywords Baja California, Ctenosaura conspicuosa, Ctenosaura hemilopha, Ctenosaura nolascensis, genealogy, human-mediated dispersal, Iguanidae, island biogeography, molecular clock, reptiles doi: /j x

2 Human translocation is not responsible for Ctenosaura hemilopha dispersal INTRODUCTION Untangling the historical causes of the geographic distributions of species or species complexes is central to the study of biogeography. However, the factors that are responsible for current species distributions cannot always be directly inferred from available data. For example, historical climatic conditions are often difficult to determine, although these may explain current species ranges and distributions. Likewise, the geological history of many areas is not completely understood; it may not be written in stone (e.g. Murphy & Aguirre-León, 2002; Riddle et al., 2008). Human impacts on species distributions must also be considered, because humans have both deliberately and accidentally mediated the dispersal of many plants and animals (e.g. Austin, 1999; Nabhan, 2002; Carlton, 2003). In this study we use molecular techniques to investigate whether or not human-mediated translocations played a role in the evolutionary history of an iguanid species complex in the Sea of Cortés. The Cape spiny-tailed iguana (Ctenosaura hemilopha Cope, 1863) species complex (Squamata, Iguanidae) is found in the southern part of the Baja Californian peninsula, on mainland Mexico (Sonora and Sinaloa), and on several islands in the Sea of Cortés (Smith, 1935; Lowe & Norris, 1955; Fig. 1). Colour pattern variability and other morphological attributes among individuals from these isolated locations led Smith (1972) to recognize five geographically isolated subspecies, four of which Grismer (1999) elevated to full species level. Ctenosaura hemilopha occurs on the southern half of the peninsula of Baja California, and C. h. insulana is found on Isla Cerralvo, c km from the peninsula (Murphy et al., 2002). Ctenosaura macrolopha (Grismer, 1999) is found on the Mexican mainland, from Hermosillo, Sonora, southwards to mid- Sinaloa. Ctenosaura nolascensis (Grismer, 1999) is restricted to Isla San Pedro Nolasco, a small island c km off the coast near Guaymas, Sonora (Murphy et al., 2002). Finally, Ctenosaura conspicuosa (Grismer, 1999) occurs only on Isla San Esteban and the neighbouring Isla Cholludo (also referred to as Isla Lobos, e.g. Smith, 1972). Isla Cholludo is located near the southernmost point of Isla Tiburón (Fig. 1), and it was connected to this island and mainland Mexico at times of maximum glaciation in the Pleistocene. The oceanic islands in the Sea of Cortés are estimated to have uplifted between 5 and 2 Ma (Carreño & Helenes, 2002). Some of these islands have never been connected to the mainland, and Isla San Esteban may be one of these (Carreño & Helenes, 2002). Thus, the occurrence of C. conspicuosa on Isla San Esteban, and the distribution of this species complex in general, is a biogeographical conundrum. Bailey (1928) suggested that C. conspicuosa on Isla San Esteban were in all probability carried there by man. When Smith (1972) later described the taxon, he suggested that the insular populations were founded by individuals from the peninsula of Baja California, as a few waif populations in a 32 o 30 o see inset N km o 7 Figure 1 Distribution of the Ctenosaura hemilopha species complex, with arrows indicating Isla San Esteban and Isla San Pedro Nolasco. The location of Isla Cholludo is indicated in the inset. Numbers indicate samples included in the analysis (see Table 1 for sample details). Shaded areas indicate the ranges of C. hemilopha (dotted line) and C. macrolopha (dashed line). Samples 1 6, C. hemilopha; 13, 14, 20 24, C. conspicuosa; 7, 18, 19, 25 28, C. macrolopha; 29 36, C. nolascensis. 26 o 24 o 29.7 o 13, Isla Tiburon Isla Cholludo Isla Datil 29.6 o Isla San Esteban km o o o o 117 o 115 o 113 o 111 o 109 o 107 o , 3 18, 19 Journal of Biogeography 38,

3 C. M. Davy et al. sweepstake pattern reached a number of the Gulf Islands. Grismer (1994, 2002) further considered the hypothesis that the indigenous culture in and around the Sea of Cortés mediated the dispersal of Ctenosaura sp. (presumably C. nolascensis) from Isla San Pedro Nolasco to Isla San Esteban. Nabhan (2003) documented in detail the complex cultural relationship between the Seri (Comcáac) people indigenous to the Sea of Cortés and the native reptiles of the region. Many Seri recognize snakes, lizards, tortoises and marine turtles by species, and some species have more than one common name in the Comcáac language. Each species has a cultural significance to the Seri. Some may be included in feasts at important celebrations; for example, marine turtles are served during coming-of-age ceremonies. Some are avoided; for example, the Seri believe that looking at certain lizards can cause a pregnant woman to miscarry (Nabhan, 2003). Along with the cultural importance of the Seri s relationship with reptiles, Seri oral history contains information about historical translocations of reptiles. Nabhan (2002, 2003) documented Seri accounts of the deliberate translocation of chuckwallas (Sauromalus) between islands in the Sea of Cortés. Chuckwallas are an important source of food for the Seri (Nabhan, 2003). Both molecular evidence and Seri traditional knowledge suggest that the Seri were responsible for translocating Sauromalus hispidus from Isla Ángel de la Guarda southwards to Isla San Lorenzo Sur, and probably also to Isla San Lorenzo Norte and Islote Granito (Petren & Case, 1997, 2002; Murphy & Aguirre-León, 2002). Seri involvement is also implicated in the dispersal of several other reptilian species throughout the Gulf islands, including side-blotched lizards, which probably dispersed as hitchhikers (Uta; Upton & Murphy, 1997), and giant chuckwallas (Sauromalus varius; Murphy & Aguirre-León, 2002), which were probably translocated deliberately (Nabhan, 2003). Human translocations have also mediated the dispersal and subsequent divergence of lizards in other parts of the world. For example, molecular data demonstrate how Lipinia noctua took the express train to distant Polynesian islands by hitchhiking with humans dispersing out of Melanesia (Austin, 1999). Could translocations by the Seri explain the peculiar occurrence of C. conspicuosa on the islands of San Esteban and Cholludo, which are so far north of the other insular Ctenosaura and surrounded by islands on which Ctenosaura are not found? The Seri people hunt spiny-tailed iguanas (Nabhan, 2003), so it would have benefited them to move Ctenosaura species to islands on which they lived or hunted. They have successfully translocated and established new populations of other iguanid lizards, as evidenced by their translocations of Sauromalus and other reptiles, and they have an oral history of the translocation of C. conspicuosa from Isla San Esteban to nearby Isla Cholludo. However, Isla San Esteban itself is located far to the north of the other insular species of Ctenosaura (Fig. 1) and is isolated from other populations of Ctenosaura. The occurrence of C. conspicuosa on Isla San Esteban is, therefore, more difficult to explain. When Nabhan (2003) directly asked a Seri elder if his people had translocated Ctenosaura from Isla San Pedro Nolasco to Isla San Esteban, the response was that, although it was certainly possible, they had no history of such a translocation. When translocating animals for live food, the Seri have a practice of breaking the legs of lizards in order to prevent escape (Nabhan, 2003), which makes accidental introductions unlikely (although not impossible). Recent translocations are unlikely to be detectable morphologically. For example, translocated populations of chuckwallas cannot be morphologically distinguished based on their place of origin. In contrast, phenotypic distinctions between C. conspicuosa on Isla San Esteban and C. nolascensis on Isla San Pedro Nolasco have been listed by Grismer (1999). These include the presence of small black spots on the ventral surface of the hind limbs of adult C. nolascensis, while C. conspicuosa, C. macrolopha and C. hemilopha have large circular blotches. In C. nolascensis, the dorsal hind limb pattern is mottled, while in C. conspicuosa it is banded. Hatchling coloration tends to differ between the populations as well, although less consistently than adult coloration (Grismer, 1999). Consistent differences in coloration between these two populations (Smith, 1972; Grismer, 1999) imply prolonged reproductive isolation, which is inconsistent with ongoing human translocation of individuals between recently diverged populations. Overall, the evidence for humanmediated dispersal of Ctenosaura to Isla San Esteban is currently equivocal. In the case of Sauromalus, molecular evidence helped to confirm Seri translocations, but such evidence is lacking for the C. hemilopha complex. To date, the sole genetic analysis of the C. hemilopha species complex, an MSc thesis (Cryder, 1999), suggested a genealogy based on 22 cytochrome b (cyt b) and cytochrome c oxidase subunit III (COIII) sequences (Fig. 2). Grismer (2002) cited Cryder s genealogy and the morphological variation between species (Grismer, 1999) as evidence that the Seri people had created C. conspicuosa by moving C. nolascensis from Isla San Pedro Nolasco to Isla San Esteban. Grismer (2002) also cited Nabhan s (2003) ethno-herpetological study of the Seri culture as evidence for Seri translocation of Ctenosaura from Isla San Pedro Nolasco to Isla San Esteban. The question of Seri involvement has not yet been adequately addressed. The Seri elders cited by Grismer (2002) did not, in fact, claim an oral history of Ctenosaura translocations between those two islands (Nabhan, 2003). The practice of breaking the legs of lizards being transported for food makes accidental translocations unlikely. Equally problematic, Cryder s (1999) phylogenetic results (Fig. 2) do not show the topology expected under the hypothesis of the translocation of Ctenosaura from Isla San Pedro Nolasco to Isla San Esteban. Recent divergence between young species or diverging populations typically manifests in genealogies as incomplete lineage sorting (e.g. Murphy & Aguirre-León, 2002; Morando et al., 2004; Heckman et al., 2007). In contrast to this prediction, Cryder s (1999) genealogy showed that the four species and the one subspecies form distinct lineages. The exception was C. nolascensis, which had two independent 274 Journal of Biogeography 38,

4 Human translocation is not responsible for Ctenosaura hemilopha dispersal C. nolascensis Isla San Pedro Nolasco C. conspicuosa Isla San Esteban C. conspicuosa Isla Cholludo C. macrolopha Sonora C. nolascensis Isla San Pedro Nolasco Figure 2 Genealogy of the matrilines of the Ctenosaura hemilopha species complex inferred by Cryder (1999) from cytochrome b and cytochrome c oxidase subunit III sequences using maximum parsimony methods (redrawn from Grismer, 2002). C. hemilopha C. hemilopha Baja California Isla Cerralvo maternal lineages, both of which were resolved and neither of which nested within another species. We used mtdna sequences to test the hypothesis suggested by Bailey (1928) and Grismer (2002) that the Seri people founded the population of C. conspicuosa on Isla San Esteban by translocating C. nolascensis from Isla San Pedro Nolasco. Acceptance of this hypothesis requires that the divergence of the separate species lineages occurred after the first known human colonization of the Americas (c. 16,500 years ago; Goebel et al., 2008). We inferred the relationships between maternal lineages of the C. hemilopha species complex using standard phylogenetic methods, and used Bayesian inference (BI) to estimate divergence times between the species. Rejection of the hypothesis requires that estimated divergence times between the species occurred before human colonization of the Americas, and that there is no evidence of incomplete lineage sorting between matrilines sampled within the ranges of C. conspicuosa and C. nolascensis. MATERIALS AND METHODS Because the sequences obtained by Cryder (1999) were not available in GenBank, we resampled the four species. Phylogenetic analysis of mtdna was assumed to produce a genealogy of maternal lineages that closely reflects the evolutionary history of the species in question (e.g. Upton & Murphy, 1997). To produce a mtdna genealogy for the C. hemilopha complex, we examined mitochondrial cyt b and COIII sequences from 31 individual Ctenosaura representing the four recognized species in the C. hemilopha complex. Where possible, we attempted to sample from a number of locations within the range of each species in order to avoid a geographic sampling bias. The subspecies C. h. insulana was not included owing to sampling restrictions. We also collected tissues from the Mexican spinytailed iguana (C. pectinata Wiegmann) in Sinaloa (near Chametla, Culiacán and Mazatlán) as an outgroup taxon. Previously, this iguanid was shown to be closely related to the C. hemilopha complex (Köhler et al., 2000). As more distant outgroups we included sequences from Petrosaurus thalassinus Cope collected in Baja California, Iguana iguana Linnaeus taken from GenBank, and Sauromalus ater Duméril collected from Sonora (near Sonoyta and Caborca). Petrosaurus thalassinus was specified as the most distant outgroup whenever required. GenBank accession numbers and voucher specimen information for all individuals are listed in Table 1. DNA extraction, amplification and sequencing We isolated total genomic DNA from frozen or 95% ethanolpreserved tissues using standard proteinase K digestion followed by phenol-chloroform extraction (Sambrook et al., 1989). Cyt b and COIII were amplified using polymerase chain reaction (PCR) (Saiki et al., 1988). DNA amplification and purification followed the methods of Blair et al. (2009), using the primers and primer-specific annealing temperatures listed in Table 2. Sequencing reactions were performed on a Gene- Amp 9700 thermal cycler (Applied Biosystems, Foster City, CA, USA), using the BigDye Terminator v 3.1 Cycle Sequencing kit (Applied Biosystems). Sequences were visualized on an ABI 377 automated sequencer (Applied Biosystems). Journal of Biogeography 38,

5 C. M. Davy et al. Species Sample ID Voucher number/field number GenBank accession no. (cyt b) GenBank accession no. (COIII) Ctenosaura hemilopha CH1 ROM HQ HQ CH2 RWM 2280 HQ HQ CH3 RWM 2282 HQ HQ CH4 RWM 623 HQ HQ CH5 RWM 631 HQ HQ CH6 RWM 879 HQ HQ C. macrolopha CH7 JRO 645 HQ HQ CH18 ROM HQ HQ CH19 ROM HQ HQ CH25 ROM HQ HQ CH26 ROM HQ HQ CH27 ROM HQ HQ CH28 ROM HQ HQ CH37 ROM HQ HQ CH38 ROM HQ HQ CH39 ROM HQ HQ C. conspicuosa CH13 KP-EC102 HQ HQ CH14 KP-EC103 HQ HQ CH 20 ROM HQ HQ CH21 ROM HQ HQ CH22 ROM HQ HQ CH23 ROM HQ HQ CH24 ROM HQ HQ C. nolascensis CH29 ROM HQ HQ CH30 ROM HQ HQ CH31 ROM HQ HQ CH32 ROM HQ HQ CH33 ROM HQ HQ CH34 ROM HQ HQ CH35 ROM HQ HQ CH36 ROM HQ HQ Petrosaurus thalassinus Petrosaurus RWM 2263 HQ HQ Iguana iguana Iguana GenBank AJ AJ Sauromalus ater S1 KP-20 HQ HQ S2 KP-22a HQ HQ S3 KP-22b HQ HQ S4 KP-19 HQ HQ S5 KP-21 HQ HQ C. pectinata CP1 ROM HQ HQ CP2 ROM HQ HQ CP3 ROM HQ HQ CP4 ROM HQ HQ CP5 ROM HQ HQ CP6 ROM HQ HQ CP7 ROM HQ HQ CP8 ROM HQ HQ CP9 ROM HQ HQ CP10 ROM HQ HQ CP11 ROM HQ HQ Table 1 Sample details for iguanid lizards included in this study. Sample ID numbers correspond to the ID numbers on the trees and in Fig. 1. Where vouchers were taken, ROM accession numbers are designated (i.e. ROM xxxx ). If vouchers were not taken, the field number corresponding to the tissue sample is indicated. Cyt b, cytochrome b; COIII, cytochrome c oxidase subunit III. Alignment and sequence analysis We sequenced 1632 base pairs combined from COIII and cyt b for 49 individuals. Sequences were aligned with ClustalW (Larkin et al., 2007) and subsequently checked by eye. Conversion to amino acids confirmed the alignment. We calculated the percentage sequence divergence (uncorrected p-distances) for all ingroup taxa based on 276 Journal of Biogeography 38,

6 Human translocation is not responsible for Ctenosaura hemilopha dispersal Table 2 Primers used to amplify and sequence cytochrome b (cyt b) and cytochrome c oxidase subunit III (COIII) from the Ctenosaura hemilopha species complex, C. pectinata, Sauromalus ater and Petrosaurus thalassinus. Target gene Primer/annealing temperature Sequence (5 3 ) Source Cyt b GLUDG-L TGACTTGAARAACCAYCGTTG Palumbi et al. (1991) 50 C CTEN-8H TTACTGTGGCGCCTCGGAAGGATATTTGGCCTCA Cryder (1999) 50 C COIII L8618CO3 CATGATAACACATAATGACCCACCAA Cryder (1999) 46 C H9323CO3 46 C ACTACGTCTACGAAATGTCAGTATCA Cryder (1999) Petrosaurus and Sauromalus cyt b Cyt b B1L CCATCCAACATCTCAGCATGATGAAA Kocher et al. (1989) Cyt b B6H 50 C GTCTTCAGTTTTTGGTTTACAAGAC Tim Birt (Queens University, pers. comm.) uncorrected p-distances as implemented in mega 4 (Tamura et al., 2007). Because direct comparison of our sampled sequences with those of Cryder (1999) was not possible, we initially followed his methods in order to determine if any inherent differences between the two data sets might have caused our genealogy to differ in topology from his (Fig. 2). Thus, maximum parsimony (MP) analysis was implemented in paup 4.0b10 (Swofford, 2002), employing a heuristic search with 50 random addition sequences (RAS) and tree bisection reconnection branch swapping. We then assessed nodal confidence for the MP strict consensus tree by nonparametric bootstrapping (Felsenstein, 1985) using a heuristic search with 1000 pseudoreplicates, 50 RAS per pseudoreplicate, and nearestneighbour interchange branch swapping (Nei & Kumar, 2000). We considered bootstrap values > 70 to indicate strong nodal support (Hillis & Bull, 1993). We used MrBayes (Huelsenbeck & Ronquist, 2001; Ronquist & Huelsenbeck, 2003) to determine the most probable evolutionary history for the matrilines based on the available sequences. MrModeltest 2.3 (Nylander, 2004) indicated that the best-fit evolutionary model for our data was (GTR+I+C), which was selected for each gene using the Akaike information criterion (Akaike, 1974, 1979). To account for potential rate variation and differing rates of substitution between the two genes (Nylander, 2004; Brandley et al., 2005), we partitioned our data set by gene, and set the analysis to account for variable rates between partitions using the command prset applyto = (all) ratepr = variable in MrBayes. The Bayesian analysis was run for generations, with two simultaneous runs of six chains sampled at 100-generation intervals. The first 2500 trees (25%) were discarded as burn-in, and the inferred genealogy was based on 7500 data points. Examination of the raw trace, the log-likelihood plot and the standard deviations of the split frequencies all indicated that convergence had occurred, and that the burn-in period was sufficient. We considered lineages to have significant support if they had posterior probability values 0.95 (Felsenstein, 2004). Estimates of divergence time It is not advisable to estimate divergence times for nodes in a tree without at least one point of geological or palaeontological reference, and preferably several (Benton & Ayala, 2003; Reisz & Müller, 2004). Robust estimations of divergence dates require several points of reference, preferably from fossil evidence. Unfortunately, the fossil record for iguanid lizards in western Mexico is scarce, and it is therefore not possible to date the evolution of the C. hemilopha species complex by dating fossils of these species. Therefore, we base our estimates of divergence time within the C. hemilopha complex on the divergence time (most recent common ancestor, MRCA) of C. pectinata and C. hemilopha estimated by Zarza et al. (2008). Using a Bayesian approach and calibration points based on fossil evidence, they estimate that the divergence of C. pectinata and C. hemilopha occurred 9.24 Ma (SD = 2.9), and that divergence within C. pectinata began between 2.3 and 6.5 Ma (Zarza et al., 2008). We use these priors in our analysis. We used the program beast (Drummond et al., 2007) to infer divergence times for species of the C. hemilopha complex under an uncorrelated lognormal relaxed molecular clock model (Drummond et al., 2006). Input files were created with BEAUti (Drummond et al., 2007), and manually edited to partition the data by gene and to specify substitution rates, gamma shape parameters and proportion of invariable sites for each partition based on the estimates made in MrModeltest. We specified monophyly of the lineages identified by our BI analysis but did not specify an input tree. beast analyses considered the Yule process tree prior, as recommended for analyses of speciation (Drummond et al., 2007). Three Markov chain Monte Carlo (MCMC) runs were made, each with generations, sampling every 100 generations with a burn-in period of 10% of the samples. Journal of Biogeography 38,

7 C. M. Davy et al. We examined output files for each of the three runs in Tracer (Rambaut & Drummond, 2007) to assess whether or not they had converged on similar estimates of divergence times. Next, we combined the results of the runs using LogCombiner (Drummond et al., 2007) for further analysis, sampling the combined runs every 300 generations for a final sample of 89,991,300 states with a burn-in of 10% of the samples. We considered effective sample size (ESS) values > 200 to indicate good mixing and a valid estimate of continuous parameters and likelihoods given the specified priors (Drummond et al., 2007). We checked the distribution of the standard deviation of the uncorrelated lognormal relaxed clock model and its coefficient of variation in Tracer during examination of the output files. Neither parameter approached zero, indicating rate variation between branches and suggesting that a strict clock would have been an inappropriate model for our data. The maximum clade credibility tree for the combined runs was computed using TreeAnnotator (Drummond et al., 2007). We used the estimated lower bound of the 95% highest posterior density (HPD) region of the MRCA parameters to test the hypothesis that C. conspicuosa and C. nolascensis could have diverged as a result of human-mediated dispersal, that is, after human colonization of the Americas (c. 16,500 years ago; Goebel et al., 2008). RESULTS Sequence analysis Average percentage sequence divergences (p-distances) within and between C. pectinata and species in the C. hemilopha complex are summarized in Table 3. Sequence divergence between C. pectinata and the C. hemilopha species complex averaged 11.2%. Divergence between lineages within the C. hemilopha species complex ranged from 0.8% to 4.5%, with the highest percentage divergence occurring between two lineages in C. nolascensis. Of 1632 nucleotide sites, 547 were variable and 378 were potentially phylogenetically informative. Within the sequences from C. hemilopha and C. pectinata (excluding other outgroups), 241 sites were potentially phylogenetically informative. The MP analysis recovered 2111 most-parsimonious trees of 925 steps (consistency index = 0.765, retention index = 0.943). The topologies of the MP strict consensus tree and the BI majority-rule consensus tree differed slightly at the tips. The trees also differed in their placement of Ctenosaura in relation to the three outgroup genera, but both methods inferred the same relationships between the five species of Ctenosaura, and resolved the same major lineages within Ctenosaura, without exception (Fig. 3). Six major mtdna lineages were recovered. Ctenosaura pectinata formed a distinct lineage distantly related to the C. hemilopha complex. Ctenosaura nolascensis was resolved into two distinct and distantly related lineages. The first of these lineages (C. nolascensis-1) was resolved as sister to a single sample of C. macrolopha from Culiacán, Sinaloa, and this group was resolved as sister to C. conspicuosa from Isla San Esteban. The remaining samples of C. nolascensis formed a separate lineage (C. nolascensis-2), sister to a lineage containing both C. hemilopha and C. macrolopha. There was high bootstrap and posterior probability support for most nodes between species (Fig. 3). Our genealogy recovered the same mtdna groups as Cryder (1999), but differed slightly in the relationships between C. nolascensis-2, C. hemilopha and C. macrolopha. Otherwise, our genealogies agreed on the relationship between the matrilines. Interestingly, the proportion of the two C. nolascensis haplotypes in our sample (5:3) approximates that shown in Fig. 2 (3:2; Cryder, 1999). Based on the sample from Culiacán (CH19), which was resolved as sister to the C. nolascensis-1 lineage, C. macrolopha did not consist of a single matrilineal lineage but contained at least two distinct matrilines. Finally, the tree topology and p-distances indicated that the degree of divergence between C. conspicuosa and C. nolascensis-1 was comparable to the divergence between C. macrolopha and C. hemilopha (Table 4). Estimates of divergence time Estimated divergence times within the major lineages and the 95% HPD of the estimates are summarized in Table 4. Estimated divergence times for all major nodes are also shown on the maximum clade credibility tree generated by TreeAnnotator from the three combined MCMC runs (Fig. 4). The estimated divergence time for the matrilines within C. conspicuosa was 1.73 Ma, with a lower 95% HPD of 326,200 years ago. Divergence between the C. nolascensis-1 Table 3 Average pairwise genetic divergence (percentage uncorrected p-distances) within and between Ctenosaura pectinata, C. hemilopha, C. macrolopha, C. conspicuosa and C. nolascensis. The two matrilines within C. nolascensis are presented separately. The standard error of percentage divergence is indicated in parentheses. C. pectinata C. conspicuosa C. nolascensis-1 C. nolascensis-2 C. hemilopha C. macrolopha C. pectinata 0.9 (± 0.2) C. conspicuosa 11.1 (± 1.2) 0.1 (± 0.1) C. nolascensis (± 1.2) 0.8 (± 0.2) 0.0 (± 0.1) C. nolascensis (± 0.012) 4.0 (± 0.6) 4.5 (± 0.2) 0.0 (± 0.0) C. hemilopha 11.3 (± 1.2) 3.9 (± 0.6) 4.0 (± 0.6) 1.1 (± 0.3) 0.0 (± 0.0) C. macrolopha 11.3 (± 1.2) 4.3 (± 0.6) 1.5 (± 0.6) 1.5 (± 0.3) 0.9(± 0.2) 0.1 (± 0.0) 278 Journal of Biogeography 38,

8 Human translocation is not responsible for Ctenosaura hemilopha dispersal Figure 3 Evolutionary history of the matrilines of the Ctenosaura hemilopha complex inferred using Bayesian inference analysis of cytochrome b and cytochrome c oxidase subunit III sequence data. C.h., Ctenosaura hemilopha complex; C.p., C. pectinata; S, Sauromalus ater. Numbers at nodes indicate Bayesian posterior probability values followed by the percentage of replicate trees in which the associated taxa clustered together in the nonparametric bootstrap analysis; indicates a posterior probability/ bootstrap proportion = 1.0/ / / / / / / /98 1.0/ /85 1.0/99 1.0/93 1.0/ / / / / P. thalassinus C.h. 19 C. macrolopha C.h. 30 C.h. 31 C.h. 32 C. nolascensis-1 C.h. 35 C.h. 36 C.h. 13 C.h. 20 C.h. 14 C.h. 21 C. conspicuosa C.h. 24 C.h. 22 C.h. 23 C.h. 1 C.h. 3 C.h. 4 C. hemilopha C.h. 5 C.h. 6 C.h. 2 C.h. 37 C.h. 38 C.h. 39 C.h. 26 C. macrolopha C.h. 7 C.h. 25 C.h. 27 C.h. 28 C.h. 18 C.h. 29 C.h. 33 C. nolascensis-2 C.h. 34 C.p. 3 C.p. 4 C.p. 1 C.p. 2 C.p. 11 C. pectinata C.p. 5 C.p. 6 C.p. 7 C.p. 10 C.p. 8 C.p. 9 S4 S2 S3 S5 S1 I. iguana Table 4 Estimated dates of divergence from the most recent common ancestor (MRCA) between the major matrilines within the Ctenosaura hemilopha species complex. Dates (Ma) were estimated under an uncorrelated relaxed clock model in beast, with the analysis partitioned by gene and priors as described in the Materials and Methods. Mean estimated divergence dates are listed, with the upper and lower bounds of the 95% highest posterior density (HPD) of each estimate. MRCA Mean (Ma) 95% HPD upper 95% HPD lower Between lineages (hemilopha/macrolopha) nolascensis-2 hemilopha + macrolopha conspicuosa + nolascensis nolascensis-1 + CH (C. macrolopha) Within lineages pectinata hemilopha macrolopha conspicuosa nolascensis nolascensis Excluding the sample of C. macrolopha from Culiacán (CH19), which was resolved as sister to this lineage. Journal of Biogeography 38,

9 C. M. Davy et al C.h C.h. 32 C.h C.h. 30 C.h. 31 C.h C.h. 14 C.h. 20 C.h. 13 C.h C.h. 24 C.h. 21 C.h. 22 C.h. 2 C.h C.h. 1 C.h. 5 C.h. 6 C.h. 3 C.h. 7 C.h C.h. 28 C.h. 25 C.h. 27 C.h C.h. 18 C.h C.h C.h C.h. 34 C.h C.p C.p C.p. 10 C.p C.p C.p. 1 C.p. 4 C.p. 3 C.p C.p. 8 C.p. 9 I. iguana C. macrolopha C. nolascensis-1 C. conspicuosa C. hemilopha C. macrolopha C. nolascensis-2 C. pectinata Figure 4 Chronogram of the Ctenosaura hemilopha species complex: maximum clade credibility tree from three combined Markov chain Monte Carlo analyses performed in beast. Ctenosaura pectinata and Iguana iguana are included as outgroups. Posterior estimates of divergence times were inferred by partitioning analyses by gene, and placing constraints on the divergence dates of two nodes (see Materials and Methods). Values at nodes indicate posterior mean ages (Ma), and node bars represent the 95% highest probability density (HPD). indicates Bayesian posterior probabilities > 95%. The full extent of the upper 95% HPD for the two most basal nodes is not shown. The scale bar shows time in millions of years ago. and C. conspicuosa matrilines was estimated at 2.9 Ma ( Ma). The most recent divergence occurred between the matrilines within C. nolascensis-1 ( Ma). DISCUSSION Genealogies for the C. hemilopha complex and estimates of divergence time between the species in the complex suggest that historical human involvement in their divergence is highly unlikely. We cannot refute the null hypothesis that C. conspicuosa diverged from the other taxa in the C. hemilopha complex before humans arrived in North America, and we cannot accept Seri translocation as the explanation for the presence of C. conspicuosa on Isla San Esteban. Although this conclusion differs from some previous interpretations, it is in agreement with Seri traditional knowledge (Nabhan, 2002, 2003). Ctenosaura hemilopha, C. macrolopha, C. nolascensis and C. conspicuosa show no evidence of recent female dispersal and gene flow between them, although there are interesting patterns present in the genealogy, as we discuss below. The occurrence of a C. nolascensis-like haplotype on the mainland is especially intriguing and suggests historical dispersal between Isla San Pedro Nolasco and the mainland. Our conclusion leaves us in need of a new biogeographical explanation for the distribution of the C. hemilopha species complex. No express train for C. conspicuosa The hypothesis that Seri translocations caused the initial divergence between the insular species of Ctenosaura requires the divergence of the matrilines found on the two islands to occur after c. 16,500 years ago (Goebel et al., 2008), but the estimated divergence time between the insular matrilines of C. conspicuosa and C. nolascensis-1 is, at a minimum, 0.84 Ma (Table 4). Consequently, the molecular data do not support the theory that the Seri (or any other human culture) mediated the initial divergence of C. nolascensis and C. conspicuosa. This finding is unlikely to surprise the Seri, whose oral histories do not include such a translocation (Nabhan, 2003). As discussed earlier, the Seri s reptilian translocations show deliberation and planning. Details of other translocations 280 Journal of Biogeography 38,

10 Human translocation is not responsible for Ctenosaura hemilopha dispersal (including those of C. conspicuosa between Isla San Esteban and Isla Cholludo) recorded by Nabhan suggest that such events have been well documented in their oral history. Furthermore, the Seri practice of breaking the legs of lizards being transported for food (Nabhan, 2002) makes accidental translocations unlikely. Iguanas with broken or dislocated legs would be unlikely to escape successfully. Even if they did escape, an injured male would have difficulty mating, and crippled females might be unable to dig suitable nests. Thus, they would be unlikely to contribute to the gene pool. Legbreaking was not used during deliberate translocations of animals because the desired outcome (establishment of a new population that subsequently could be harvested) would be thwarted (Nabhan, 2002). Other incidental evidence also suggests pre-seri divergence of C. conspicuosa and C. nolascensis. The population of C. conspicuosa on Isla Cholludo founded by the Seri (Grismer, 2002; Nabhan, 2002) can be used (albeit cautiously) as a null model for the expected genealogical pattern (i.e. incomplete lineage sorting) caused by Seri translocation. Although we could not include these sequences in our analysis, Cryder s (1999) analysis included two individuals from this population, both of which fall undifferentiated into the lineage containing samples from Isla San Esteban (Fig. 2). In contrast to these patterns, the genealogical distinctiveness of populations on Isla San Pedro Nolasco and Isla San Esteban provide further evidence that these populations are not the result of human translocations. The oral history of the Seri is a valuable cultural resource, not only for the Seri themselves, but also for the rest of humanity. As such, Nabhan s (2002, 2003) documentation of the ethno-herpetology of this culture provides an important record of Seri traditional knowledge, and a relatively unique anthropological work. Unfortunately, several species of reptiles pictured in the book are misidentified. For example, Dermachelys coriacea is labelled as Caretta caretta; Pituophis is labelled as Lampropeltis; a desert tortoise (Gopherus agassizii) is described as a turtle ; a Sauromalus is labelled as Ctenosaura, with the location of the photo incorrectly listed as Isla San Esteban; and a Seri carving of a Ctenosaura is misidentified as Sauromalus (Nabhan, 2003). These misidentifications raise the question of whether other species described by Seri elders could also be misidentified (i.e. assigned an incorrect scientific name). This is not the case. The names associated with the images are those in the photo archives of the Arizona-Sonora Desert Museum, and the errors are editorial in nature (G.P. Nabhan, University of Arizona, pers. comm. to R.W.M., 14 June 2010). The discrepancy in names does not invalidate the traditional knowledge that Nabhan (2002, 2003) has so carefully collected. Along with its inherent value, traditional knowledge can be an important source of scientific inspiration, and can inform the development of scientific hypotheses and management plans (e.g. Kimmins, 2008). However, the sharing of traditional knowledge by indigenous communities is a gesture of trust, and selective interpretations of traditional knowledge by members of the scientific community can damage that trust. We acknowledge that the occurrence of C. conspicuosa on a small, oceanic island far from any obvious founder populations is a biogeographical conundrum. Human (specifically Seri Indian) activities are known to have strongly shaped the biogeography of the Sea of Cortés, and human-mediated dispersal of C. hemilopha between the islands was first suggested to explain this puzzle nearly a century ago (Bailey, 1928). There is no doubt that the Seri are capable of successfully founding insular populations of iguanid lizards (Petren & Case, 1997, 2002; Murphy & Aguirre-León, 2002; Nabhan, 2002, 2003), so Bailey s original hypothesis of humanmediated dispersal was a reasonable one. However, the current question is not one of ability, but simply of whether or not humans actually translocated C. hemilopha from Isla San Pedro Nolasco to Isla San Esteban. At this time, traditional knowledge and molecular biology both reject the hypothesis of human-mediated dispersal of C. hemilopha between these two islands, and a new interpretation of the biogeography of the C. hemilopha complex is needed. Biogeography of the C. hemilopha species complex Our tree topology differs slightly from that shown in Fig. 2 (Cryder, 1999), but both genealogies indicate that at least two independent colonization events were involved in the history of C. nolascensis on Isla San Pedro Nolasco: one by an ancestor shared with C. conspicuosa, and the other by an ancestor shared with the mainland and peninsular species. Cryder s tree places C. nolascensis-2 as the immediate sister of C. macrolopha, while our analyses resolves C. macrolopha and C. hemilopha in a lineage sister to C. nolascensis-2. It is possible that we sampled different haplotypes within C. nolascensis, which would suggest the potential presence of three or more matrilines on Isla San Pedro Nolasco. However, because the proportions of samples falling into each C. nolascensis lineage are roughly equal in the two studies, it is more likely that we sampled the same two matrilines of C. nolascensis as Cryder. Although sampling was limited in both studies, the frequencies of the two haplotypes are about 60% C. nolascensis-1:40% C. nolascensis-2. These two haplogroups had 4.5% divergence between them, a higher divergence than occurred between any two recognized species within the complex. Such genealogical patterns are often interpreted as evidence for cryptic speciation (e.g. Tavares & Baker, 2008). However, C. nolascensis is morphologically distinct from the other species of Ctenosaura (Grismer, 1999), and we know of no previous mention of distinct morphotypes within C. nolascensis on Isla San Pedro Nolasco. Because of the morphological distinctness of C. nolascensis and the lack of obvious mechanisms of reproductive isolation on the island, we find the hypothesis of sympatric cryptic species within C. nolascensis to be improbable. Differences between Cryder s and our trees probably result from different sampling strategies. Cryder s analysis included the subspecies C. h. insulana from Isla Cerralvo, which may have affected the inferred relationships among the species. Journal of Biogeography 38,

11 C. M. Davy et al. Furthermore, Cryder s (1999) analysis of C. macrolopha was based on individuals collected at a single locality within the species wide range. Because this strategy was unlikely to sample genetic diversity in mainland species, we collected samples throughout the range of C. macrolopha along the coast of Sonora and Sinaloa from the north to south extremes, as well as for about half of the range for C. hemilopha on the peninsula of Baja California (Fig. 1). However, our sampling did not fully cover these species ranges, and sampling further inland in Sonora and Sinaloa and further north on the peninsula might have discovered additional lineages. Further sampling of C. macrolopha could be extremely informative, especially because the matrilines within this species do not form a single lineage, and the species may contain several more divergent lineages. One sample of C. macrolopha (CH19, from Culiacán, Sinaloa) was resolved as sister to C. nolascensis-1, and this is particularly intriguing. The sample site is not near Isla San Pedro Nolasco (Fig. 1), and the haplotype occurred alongside the more common haplotype in C. macrolopha. Thus, there is no reason to suspect that C. macrolopha haplotypes segregate geographically. There are several potential explanations for the placement of this sample in the genealogy. First, it could be the result of incomplete lineage sorting. Second, the pattern could be caused by Seri (or other human) transport of C. nolascensis to the mainland, followed by deliberate or accidental introduction into the wild. Given the estimated average divergence date of 1.57 Ma for this haplotype (95% HPD = Ma; Fig. 4), we find these two explanations unlikely. Third, this pattern could indicate historical dispersal from San Pedro Nolasco to the mainland, followed by cytoplasmic capture and incorporation of the C. nolascensis-like haplotype into C. macrolopha. The presence of iguanids on many deep-water oceanic islands testifies to their ability to disperse across oceans, and Isla San Pedro Nolasco is only 14.6 km offshore. More rigorous sampling of the matrilines present throughout Sonora and Sinaloa may shed further light on the question. The most perplexing aspect of the data is the counterintuitive pattern of divergence, whereby the mainland and peninsular species diverged after their origin from the two insular species. Ctenosaura conspicuosa and C. nolascensis became isolated from the mainland population of the C. hemilopha complex after their divergence from C. pectinata. The peninsula of Baja California, being formed more than 5 Ma (Murphy & Aguirre-León, 2002), is much older than the minimum estimated Pleistocene divergence between the two non-insular forms, C. hemilopha and C. macrolopha (Table 4). This leads to two possible biogeographical scenarios. Ancestral Ctenosaura from the mainland may have dispersed around the head of the Sea of Cortés into the southern part of the peninsula. Alternatively, the distribution may have been continuous, and isolation could reflect Pleistocene climatic changes. The absence of fossil Ctenosaura from California, Arizona and the peninsula of Baja California suggests that dispersal was probably involved. Regardless, isolation of the peninsular population from the mainland population led to a cladogenic event, probably during the Pleistocene. Evidence from nuclear genes would allow the construction of a more informative species phylogeny, which may clarify the evolutionary history of these species. The detailed biogeography of this species complex continues to defy detailed interpretation, but we now know that humans were not involved. Human activities have played a pivotal role in the biogeographical history of many species, but testing hypotheses of human translocations is often challenging, and limited by the available evidence. We successfully used traditional mtdna analysis, together with Bayesian methods, to refute the hypothesis of human-mediated dispersal in the case of the C. hemilopha complex, and reached an alternative conclusion that is in keeping with the traditional knowledge of the Seri people. In our case, the traditional knowledge of the Seri was carefully documented (Nabhan, 2002, 2003), but this is often not the case. The teachings of many cultures are rapidly being lost. It is our hope that the information contained in traditional teachings can be preserved, and that such knowledge will continue to inform scientific studies. ACKNOWLEDGEMENTS We especially thank the Seri people of Tiburón and Isla San Esteban for permission to enter and sample lizards on their traditional lands. Sample collection in Mexico was approved by the Mexican government (SEMARNAT SGPA/ DGVS/ 03489/07). Access to an MSc thesis from Loma Linda University was kindly granted by W. Hays; R.L. Carter made valiant attempts to retrieve M. Cryder s sequence data; and G.P. Nabhan graciously provided important information. We thank C. Blair and two anonymous referees for valuable comments on earlier versions of the manuscript. C.M.D. was supported by a Canada Graduate Scholarship from the National Research Council of Canada. This research was supported by grants from the Natural Sciences and Engineering Research Council of Canada, Discovery Grant A3148, the Royal Ontario Museum (ROM) Foundation, and the ROM Members Volunteer Committee to R.W.M. REFERENCES Akaike, H. (1974) A new look at the statistical model identification. Institute of Electrical and Electronics Engineers Transactions on Automatic Control, 19, Akaike, H. (1979) A Bayesian extension of the minimum AIC procedure of autoregressive model fitting. Biometrika, 66, Austin, C.C. (1999) Lizards took express train to Polynesia. Nature, 397, Bailey, J.W. (1928) A revision of the lizards of the genus Ctenosaura. Proceedings of the United States National Museum, 73, article 12, Benton, M.J. & Ayala, F.J. (2003) Dating the tree of life. Science, 300, Journal of Biogeography 38,

12 Human translocation is not responsible for Ctenosaura hemilopha dispersal Blair, C., Méndez de la Cruz, F.R., Ngo, A., Lindell, J., Lathrop, A. & Murphy, R.W. (2009) Molecular phylogenetics and taxonomy of leaf-toed geckos (Phyllodactylidae: Phyllodactylus) inhabiting the peninsula of Baja California. Zootaxa, 2027, Brandley, M.C., Schmitz, A. & Reeder, T.W. (2005) Partitioned Bayesian analysis, partition choice, and the phylogenetic relationships of scincid lizards. Systematic Biology, 54, Carlton, J.T. (2003) Community assembly and historical biogeography in the North Atlantic Ocean: the potential role of human-mediated dispersal vectors. Hydrobiologia, 503, 1 8. Carreño, A.L. & Helenes, J. (2002) Geology and ages of the islands. A new island biogeography of the Sea of Cortés (ed. by T.J. Case, M.L. Cody and E. Ezcurra), pp Oxford University Press, New York. Cryder, M. (1999) Molecular systematics and evolution of the Ctenosaura hemilopha complex (Squamata; Iguanidae). MSc Thesis, Loma Linda University, Loma Linda, CA. Drummond, A.J., Ho, S.Y.W., Phillips, M.J. & Rambaut, A. (2006) Relaxed phylogenetics and dating with confidence. PLoS Biology, 4, Drummond, A.J., Ho, S.Y.W., Philips, M.J. & Rambaut, A. (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology, 7, Felsenstein, J. (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution, 39, Felsenstein, J. (2004) Inferring phylogenies. Sinauer Associates, Sunderland, MA. Goebel, T., Waters, M.R. & O Rourke, D.H. (2008) The late Pleistocene dispersal of modern humans in the Americas. Science, 319, Grismer, L.L. (1994) Geographic origins for the reptiles on islands in the Sea of Cortés, México. Herpetological Natural History, 2, Grismer, L.L. (1999) An evolutionary classification of reptiles on islands in the Sea of Cortés, México. Herpetologica, 55, Grismer, L.L. (2002) Spiny-tailed iguanas, insular evolution, and Seri Indians: how long does it take to make a new species and does it matter who makes it? Iguana Times, 9, 3 8. Heckman, K.L., Mariani, C.L., Rasoloarison, R. & Yoder, A.D. (2007) Multiple nuclear loci reveal patterns of incomplete lineage sorting and complex species history within western mouse lemurs (Microcebus). Molecular Phylogenetics and Evolution, 43, Hillis, D.M. & Bull, J.J. (1993) An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Systematic Biology, 42, Huelsenbeck, J.P. & Ronquist, F. (2001) MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics, 17, Kimmins, J.P. (2008) From science to stewardship: harnessing forest ecology in the service of society. Forest Ecology and Management, 256, Kocher, T.D., Thomas, W.K., Meyer, A., Edwards, S.V., Paabo, S., Villablanca, F.X. & Wilson, A.C. (1989) Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences USA, 86, Köhler, G., Schroth, W. & Streit, B. (2000) Systematics of the Ctenosaura group of lizards (Reptilia: Sauria: Iguanidae). Amphibia-Reptilia, 21, Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T.J. & Higgins, D.G. (2007) ClustalW and ClustalX version 2.0. Bioinformatics, 23, Lowe, C.H., Jr & Norris, K.S. (1955) Analysis of the herpetofauna of Baja California, México. III. New and revived reptilian subspecies of Isla San Esteban, Sea of Cortés, Sonora, México, with notes on the satellite islands of Isla Tiburón. Herpetologica, 11, Morando, M., Avila, L.J., Baker, J. & Sites, J.W., Jr (2004) Phylogeny and phylogeography of the Liolaemus darwinii complex (Squamata: Liolaemidae): evidence for introgression and incomplete lineage sorting. Evolution, 58, Murphy, R.W. & Aguirre-León, G. (2002) The nonavian reptiles: origins and evolution. A new island biogeography of the Sea of Cortés (ed. by T.J. Case, M.L. Cody and E. Ezcurra), pp Oxford University Press, New York. Murphy, R.W., Sánchez-Piňero, F., Polis, G.A. & Aalbu, R.L. (2002) New measurements of area and distance for islands in the Sea of Cortés. A new island biogeography of the Sea of Cortés (ed. by T.J. Case, M.L. Cody and E. Ezcurra), pp Oxford University Press, New York. Nabhan, G.P. (2002) Cultural dispersal of plants and reptiles. A new island biogeography of the Sea of Cortés (ed. by T.J. Case, M.L. Cody and E. Ezcurra), pp Oxford University Press, New York. Nabhan, G.P. (2003) Singing the turtles to sea: the Comcáac (Seri) art and science of reptiles. University of California Press, Berkeley, CA. Nei, M. & Kumar, S. (2000) Molecular evolution and phylogenetics. Oxford University Press, New York. Nylander, J.A.A. (2004) MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University. Available at: (accessed 2 November 2008). Palumbi, S.R., Martin, A., Romano, S., McMillan, W.O., Stice, L. & Grabowski, G. (1991) The simple fool s guide to PCR, v Department of Zoology and Kewalo Marine Laboratory, University of Hawaii, Honolulu, HI. Petren, K. & Case, T.J. (1997) A phylogenetic analysis of body size evolution in chuckwallas (Sauromalus) and other iguanines. Evolution, 51, Petren, K. & Case, T.J. (2002) Updated mtdna phylogeny for Sauromalus and implications for the evolution of gigantism. A new island biogeography of the Sea of Cortés (ed. by T.J. Journal of Biogeography 38,