JOURNAL OF EXPERIMENTAL ZOOLOGY 309A (2008) A Journal of Integrative Biology Genetic Characterization of Captive Cuban Crocodiles (Crocodylus rhombifer) and Evidence of Hybridization With the American Crocodile (Crocodylus acutus) JEREMY P. WEAVER1, DAVID RODRIGUEZ1, MIRYAM VENEGAS-ANAYA1, JOSÉ ROGELIO CEDEÑO-VÁZQUEZ2, MICHAEL R.J. FORSTNER3, AND LLEWELLYN D. DENSMORE III1 1 Department of Biological Sciences, Texas Tech University, Lubbock, Texas 2 El Colegio de la Frontera Sur, Unidad Chetumal, Chetumal, Q. Roo, Me´xico 3 Department of Biology, Texas State University-San Marcos, San Marcos, Texas ABSTRACT There is a surprising lack of genetic data for the Cuban crocodile (Crocodylus rhombifer), especially given its status as a critically endangered species. Samples from captive individuals were used to genetically characterize this species in comparison with other New World crocodilians. Partial mitochondrial sequence data were generated from cyt-b (843 bp) and the tRNAPro- tRNAPhe-D-loop region (442 bp). Phylogenetic analyses were performed by generating maximum parsimony, maximum likelihood, and Bayesian-based topologies. In addition, in an effort to identify species-specific alleles, ten polymorphic microsatellite loci were genotyped. Distance and model-based clustering analyses were performed on microsatellite data, in addition to a model-based assignment of hybrid types. Both mitochondrial and nuclear markers identified two distinct C. rhombifer genetic sub-clades (a and b); and microsatellite analyses revealed that most admixed individuals were F2 hybrids between C. rhombifer-a and the American crocodile (C. acutus). All individuals in the C. rhombifer-b group were morphologically identified as C. acutus and formed a r 2008 Wiley-Liss, Inc. distinct genetic assemblage. J. Exp. Zool. 309A, 2008. How to cite this article: Weaver JP, Rodriguez D, Venegas-Anaya M, Cedeño-Vázquez JR, Forstner MRJ, Densmore III LD. 2008. Genetic characterization of captive Cuban crocodiles (Crocodylus rhombifer) and evidence of hybridization with the American crocodile (Crocodylus acutus). J. Exp. Zool. 309A:[page range]. Cuban crocodiles (Crocodylus rhombifer) are considered endangered (CITES Appendix, IUCN Red List-EN) due to their limited distribution, habitat loss, and the introduction of exotic animals into their environment (Ross, ’98). The recent encroachment of humans into C. rhombifer’s territory has limited its distributional range to about 186 square miles (300 km2) within Cienaga de Zapata and Cienaga de Lanier in southwestern Cuba (Fig. 1). Sub-fossils of C. rhombifer found in Cuba are dated to the Pleistocene (Varona, ’66; ’84); whereas several sub-fossils found on Grand Cayman (Morgan et al., ’93) and the Bahamas (Franz et al., ’95) are from the Holocene, which suggests it was historically found throughout most of the Caribbean. It has been suggested that a possible contributing factor to the decline C. rhombifer over the past 25 years may have been the introduction of the Brown caiman r 2008 WILEY-LISS, INC. (Caiman crocodilus fuscus) into Cienaga de Lanier (Ross, ’98). Although recent reports by R. Soberón (personal communication) have indicated that this is not possible because the C. rhombifer population on that island was already extirpated before C. c. fuscus was introduced. C. rhombifer is also naturally sympatric with the American crocodile (C. acutus), but C. acutus has a more extensive distribution that extends from North America into South America and the Caribbean (Ramos et al., ’94; Thorbjarnarson et al., 2006). Grant sponsor: Howard Hughes Medical Institute Grant. Correspondence to: Jeremy P. Weaver, Department of Biological Sciences, Texas Tech University, Lubbock, TX 79409. E-mail: jeremy.weaver@ttu.edu Received 16 July 2007; Revised 14 February 2008; Accepted 4 May 2008 Published online in Wiley InterScience (www.interscience.wiley. com). DOI: 10.1002/jez.471 2 WEAVER ET AL. Fig. 1. Map of Cuba with localities of C. rhombifer-like haplotypes (a and b) found outside their present range (vertical stripes). The Cuban crocodile’s breeding season overlaps with that of C. acutus by a few days in the month of January (Varona, ’66). The variation in length between these two animals is approximately 1.5 m with adult C. acutus males reaching 5 m and C. rhombifer reaching 3.5 m (Varona, ’66); thus, making hybridization physically possible. There have been several documented cases of hybridization between crocodiles in captive populations; most pertinently between C. acutus and C. rhombifer at the Laguna del Tesoro farm in Cuba (Ross, ’98), and consequently these hybrids could have been distributed to US zoological parks and private collectors. The ability to detect hybrids is essential in identifying pure breeding populations for reintroductions into extirpated areas (Allendorf et al., 2001; FitzSimmons et al., 2002). Despite the critical status of wild C. rhombifer populations there has been little or no genetic data published or reported on this species; although, there is an ongoing ecological study that will include some genetic analyses (R. Ramos and O. Sanjur, personal communication). Hybrid introgression has been detected in some New World crocodilians (Hekkala, 2004; Ray et al., 2004; Rodriguez, 2007; Cedeño-Vázquez et al., 2008; Rodriguez et al., 2008), and owing to C. rhombifer’s smaller population numbers and its frequent sympatry with C. acutus, the genetic integrity of this species is at risk. It has been suggested that detecting hybrids is less exhaustive when the two parental crocodiles possess different karyotypes, but detection of hybridization between individuals with J. Exp. Zool. similar karyotypes requires more in-depth analyses (Chavananikul et al., ’94; FitzSimmons et al., 2002). Due to the chromosomal and biochemical similarity (Cohen and Gans, ’70; Densmore, ’83) and the relatively recent divergence (Brochu, 2000) between C. rhombifer and C. acutus, detecting hybrids based on morphological characters alone may be problematic. In this specific case, the use of molecular markers is warranted. Molecular markers have been used routinely to characterize threatened species and populations (Frankham et al., 2002), but genetic studies first require a point of reference to accurately assess species assignments. Mitochondrial DNA (mtDNA) is useful in constructing phylogenies and haplotype networks. However, the maternal inheritance of mtDNA limits our ability to detect hybridization to cases where there is disagreement between morphology and mtDNA assignments. Therefore, nuclear DNA (nDNA) markers must also be utilized to more accurately evaluate species designations and determine hybrid types. Developing a pure breeding stock of C. rhombifer will be essential in maintaining the genetic integrity of the species, which is why any potential hybridization with other species can be a problem in captive populations. The purpose of our study was to genetically characterize all available samples of captive Cuban crocodiles from US zoological institutions and to describe C. rhombifer-like haplotypes found in the Caribbean, Florida, and Mexico, which may present a threat to the genetic purity of other endemic crocodile species. This work will provide the foundation for future genetic 3 GENETIC CHARACTERIZATION OF THE CUBAN CROCODILE treatments of wild Cuban crocodile populations, assist in further efforts to identify hybrids outside of Cuba, and assess the utility of genetic methods in detecting inter-specific admixture within other captive populations. MATERIALS AND METHODS Samples Wild C. rhombifer populations from Cuba could not be sampled, but zoological specimens were readily available. Whole blood or skin clips were collected from captive C. rhombifer, wild caught C. acutus, wild caught C. moreletii (Morelet’s crocodile) and a captive C. intermedius (Orinoco crocodile) (see Appendix). Blood was collected via the caudal or dorsal sinus and stored in cell lysis buffer (Gorzula et al., ’76; Bayliss, ’87), whereas skin clips were stored in 95% ethanol. Both sets of tissue were stored at 201C prior to DNA isolation. Total genomic DNA was extracted using the PureGene isolation kit (Gentra Systems, Minneapolis, MN), electrophoresed on a 2% agarose gel, and visualized with ethidium bromide under UV light. Genetic markers Mitochondrial DNA A partial cytochrome-b (cyt-b) fragment (843 bp) was amplified from C. rhombifer, C. acutus, C. moreletii, and C. intermedius using primers crCYTBfor and crCYTBrev. Primers drL15459 (modified from Glenn et al., 2002) and CR2HA (modified from Ray and Densmore, 2002) were used to amplify the tRNAPro-tRNAPhe-D-loop region (442 bp; Table 1) for only C. rhombifer and C. acutus. Polymerase chain reactions (PCR) were performed in 50 mL volumes using 0.50 mL of total genomic DNA (tDNA) (50 ng/mL), 36.25 mL of ddH20, 10 mL of buffer (0.3 M TRIS, 0.0175 M MgCl2, and 0.075 M (NH4)2SO4), 2.0 mL of 2.5 mM dNTPs, 0.50 mL (10 mM) of forward primer, 0.50 mL (10 mM) reverse of primer and 0.25 mL (1.25 U) of Promega Taq polymerase (Promega Corp., Madison, WI). Thermocycling conditions for all primers consisted of an initial denaturation step of 2 min at 941C, then 33 cycles of 30 sec at 941C, 1 min at 581C, and 45 sec at 721C; with a final extension of 7 min at 721C. Unincorporated dinucleotides and primers were removed from PCR products using the Qiagen PCR purification kit (Qiagen, Inc., Valencia, CA). Products were cycle sequenced using Big Dye v3.1 dye terminator TABLE 1. Primer sequences used to generate mtDNA fragments Primer crCYTBfor crCYTBrev CrCYTBintfor1 drL15459 CR2HA 1 Sequence 50 50 50 50 50 ATGACCCACCAACTACGAAAATC 30 CGAAGGGGTTTGATTAATAGGTT 30 TAGCAACTGCCTTCATAGGCTAC 30 AGGAAAGCGCTGGCCTTGTAA 30 GGGGCCACTAAAAACTGGGGGGA 30 Used only for sequencing. (Applied Biosystems, Inc., Foster City, CA). Cycle sequence products were purified by passing through a G-50 Sephadex column (0.5 gm of Sephadex/800 mL ddH2O), which was incubated at room temperature for 30 min and centrifuged at 3,000 rpm for 2 min to construct the column. Dried cycle sequence product was denatured in formamide and electrophoresed on an ABI 3100-Avant genetic analyzer (Applied Biosystems, Inc., Foster City, CA). Chromatograms were viewed and trimmed using Sequencher 4.1.4 (Gene Codes Corp. Ann Arbor, MI), and then aligned using Clustal X (Thompson et al., ’97) and BioEdit 5.0.6 (Hall, ’99). All newly generated sequences were accessioned into the GenBank database (EU034541-EU034627). Three sequences obtained from NCBI were used for comparison with our cyt-b sequence data, the Estuarine crocodile (C. porosus; AJ810453), the Siamese crocodile (C. siamensis; DQ353946) and the Nile crocodile (C. niloticus; AJ810452). Microsatellites Ten polymorphic microsatellite loci (Dever and Densmore, 2001; FitzSimmons et al., 2001) were also amplified for each individual (Table 2). Primers were fluorescently labeled with WellRed dyes (Beckman Coulter, Inc., Fullerton, CA) and amplified using 12.5 mL PCR reactions, which included 0.125 mL of tDNA, 9.06 mL of ddH2O, 2.50 mL buffer (0.3 M TRIS, 0.0175 M MgCl2, and 0.075 M (NH4)2SO4), 0.50 mL of 2.5 mM dNTPs (10 mM), 0.13 mL forward primer (10 mM), 0.13 mL reverse primer (10 mM), and 0.0625 mL (0.31 U) Promega Taq polymerase. PCR conditions for all primers consisted of an initial denaturation step of 2 min at 941C, then 33 cycles of 30 sec at 941C, 1 min at 581C or 621C (Table 2), and 45 sec at 721C; with a final extension of 5 min at 721C. Fragments were sized based on a 400 bp size J. Exp. Zool. 4 WEAVER ET AL. TABLE 2. Microsatellite diversity values for each genetic cluster inferred using STRUCTURE (without admixed individuals) C. rhombifer-I Locus C391 Cj16 Cj18 Cj20 Cj109 Cj119 Cj131 Cu5-123 Cud68 Cuj131 mean C. rhombifer-II C. acutus AT (C1)1 N2 A3 HO 4 HE 5 N2 A3 HO 4 HE 5 N2 A3 HO 4 HE 5 58 62 58 62 62 58 58 58 58 58 19 22 21 22 20 21 21 22 18 20 1 4 3 4 3 4 6 3 3 2 3.3 0.00 0.82 0.57 0.64 0.80 0.43 0.91 0.32 0.67 0.05 0.52 0.00 0.65 0.64 0.66 0.63 0.50 0.66 0.35 0.48 0.05 0.46 7 7 7 7 7 7 7 7 7 7 5 3 3 5 4 3 2 5 3 4 3.7 1.00 0.57 0.57 0.71 0.29 0.43 0.86 0.71 0.86 0.43 0.64 0.78 0.58 0.65 0.73 0.50 0.69 0.53 0.81 0.67 0.71 0.66 14 14 14 14 14 14 14 14 14 14 5 7 4 4 6 3 2 5 2 2 4.0 0.57 0.86 0.64 0.50 0.50 0.50 0.14 0.50 0.36 0.29 0.49 0.68 0.78 0.60 0.56 0.82 0.62 0.25 0.48 0.52 0.35 0.57 1 Annealing temperature. Number of individuals sampled per locus. 3 Number of average alleles per locus. 4 Observed heterozygosity. 5 Expected heterozygosity. 2 standard, using a CEQ8800 genetic analyzer and software (Beckman Coulter, Inc., Fullerton, CA). Using identical scoring methods, genotypes from all C. rhombifer samples were compared with 14 pure C. acutus (see Appendix) identified by Rodriguez (2007) and Rodriguez et al. (2008). Data analysis Sequence data MODELTEST (Posada and Crandall, ’98) was used to obtain the best-fit model of nucleotide substitution for maximum likelihood (ML), and MRMODELTEST (Nylander, 2004) was used to obtain the best-fit model of evolution for Bayesian inference (BI). Phylogenetic analyses of cyt-b sequence data, under the ML and maximum parsimony criteria, were performed in PAUP4.0b10 (Swofford, 2002). A starting tree was generated by stepwise addition of taxa, with swapping performed by utilizing the tree bisection reconnection algorithm. Node support was determined by bootstrapping topologies for 1,000 replications. Additionally, a BI tree, with posterior probabilities, was constructed using MrBayes v3.1.2 (Huelsenbeck and Ronquist, 2001). Four Markov chains were implemented for 1,000,000 iterations after an initial burn-in of 100,000 iterations. To obtain finer haplotype resolution, a neighbor-joining tree was constructed in PAUP using uncorrected pairwise genetic distances based on tRNAPro-tRNAPhe-D-loop sequence data. J. Exp. Zool. Node support was determined by bootstrapping the resulting topology for 1,000 iterations. Microsatellite data The program POPULATIONS v1.2.28 (Langella, ’99) was employed to estimate Dc pairwise genetic distances (Cavalli-Sforza and Edwards, ’67) to construct an exploratory neighbor-joining tree for all C. rhombifer and C. acutus individuals, which was visualized in TREEVIEW (Page, ’96). The program STRUCTURE (Pritchard et al., 2000) was used to determine assignment probabilities to specific genetic clusters by constraining K to the number of clades suggested by the neighbor-joining topology. We assumed that purebred individuals will have high assignment probabilities (40.97) to species genetic clusters, whereas hybrids will have intermediate assignment probabilities (o0.97). CERVUS 3.0 (Marshall et al., ’98) was used to estimate measures of microsatellite diversity. If hybrids were detected, then NEWHYBRIDS (Anderson and Thompson, 2002) was used to implement a Bayesian-based algorithm, which assigns individuals into six genotypic classes. Genotype classes consist of two parental groups (C. rhombifer and C. acutus), first generation hybrids (F1), second generation hybrids (F2), F1 backcrosses to C. rhombifer and F2 backcrosses to C. acutus. 5 GENETIC CHARACTERIZATION OF THE CUBAN CROCODILE RESULTS Sequence-based analyses For cyt-b sequence data, MODELTEST and MRMODELTEST indicated the best model of nucleotide substitution was GTR1G for both ML and BI. The C. rhombifer species group fell within the New World crocodilian clade that included C. niloticus (Fig. 2). Only two haplotypes were found using cyt-b sequence data (a and b), which also corresponded to two haplotypes detected using tRNAPro-tRNAPhe-D-loop sequences. Measures of percent uncorrected distances based on cyt-b sequences (Table 3), between the C. rhombifer-a haplotype and C. acutus, and between the C. rhombifer-b haplotype and C. acutus were both estimated at 5.3%. Percent divergence between the two C. rhombifer cyt-b haplotypes was 0.9%. Similarly, estimated divergences for tRNAPro-tRNAPhe-D-loop sequences were 3.4% between C. rhombifer-a haplotypes and C. acutus, and 4.1% between C. rhombifer-b haplotypes and C. acutus, respectively; whereas the genetic distance between C. rhombifer-a and C. rhombifer-b haplotypes was estimated at 1.6% (data not shown). Microsatellite analyses Fig. 2. (A) Bayesian consensus tree for cyt-b sequences, showing values (BI/MP/ML) where  5 (1.00/100/100). (B) Neighbor-joining tree based on tRNAPro-tRNAPhe-D-loop sequences with bootstrap support values. Two distinct subclades were inferred using both mitochondrial sequences. When all the samples were pooled, a matrix of Dc distances returned three distinct clades on a neighbor-joining tree (C. rhombifer-I, C. rhombifer-II, and C. acutus) with several individuals clustering between clades (Fig. 3). Using the neighbor-joining tree as a guide, K was constrained to three for model-based clustering methods implemented in STRUCTURE; these genetic clusters were named C. rhombifer-, C. rhombifer-2, and C. acutus. Posterior assignment probabilities suggested that admixture was primarily occurring between C. rhombifer-I and C. acutus (Fig. 4), therefore, individuals with intermediate probabilities were designated as hybrids (Appendix). After the STRUCTURE results, NEWHYBRIDS was used to classify hybrid types between C. rhombifer-I and C. acutus, these individuals were mostly F2 hybrids (Fig. 5). One individual, RC051, collected in Cancun, TABLE 3. Uncorrected pairwise genetic distance values for cyt-b (1) RC051 (2) MF348 (3) RC210 (4) LD175 (5) RC206 (6) AJ810452 (7) AJ810453 (8) DQ353946 (1) (2) (3) (4) (5) (6) (7) (8) C. C. C. C. C. C. C. C. rhombifer-a rhombifer-b acutus intermedius moreletii niloticus porosus siamensis – 0.009 0.053 0.057 0.052 0.079 0.104 0.121 – 0.053 0.055 0.052 0.077 0.107 0.121 – 0.015 0.045 0.056 0.093 0.107 – 0.046 0.059 0.093 0.107 – 0.061 0.098 0.110 – 0.114 0.123 – 0.021 – J. Exp. Zool. 6 WEAVER ET AL. Fig. 3. Exploratory neighbor-joining tree based on Dc distances showing microsatellite distances for C. rhombifer, C. acutus, and possible C. acutus x rhombifer hybrids (H) constructed using POPULATIONS. A model-based analysis using STRUCTURE confirmed the same species groupings and hybrid assignments (See Fig. 4). RC051 was intermediate between both C. rhombifer microsatellite clusters, and MF271(?) was resolved as a possible hybrid between C. rhombifer and C. palustris. Fig. 4. A barplot of posterior probability assignments (K constrained to 3) to species groups generated in STRUCTURE based on microsatellite data and sorted by haplotype (see Fig. 2). C.r.-II 5 C. rhombifer genetic cluster II, C.r.-I 5 C. rhombifer genetic cluster I, and C.a. 5 pure C. acutus genetic cluster. Inferred hybrids are designated by an H. J. Exp. Zool. GENETIC CHARACTERIZATION OF THE CUBAN CROCODILE 7 Fig. 5. A barplot of posterior probabilities for assignment to six genotype classes generated in NEWHYBRIDS (see text). The plot is partitioned into inferred pure species groups or hybrid types, F1 5 first filial generation, F2 5 second filial generation, C.a._Bx 5 backcross to C. acutus, and C.r.-I_Bx 5 backcross to C. rhombifer-I. Mexico was morphologically identified as C. acutus, carried a C. rhombifer-a haplotype and exhibited evidence of admixture between C. rhombifer-I and C. rhombifer-II. Captive specimen MF271, which also carried a C. rhombifer-a haplotype, was resolved as a possible hybrid between C. rhombifer and the Mugger crocodile (C. palustris), and exhibited four unique alleles (Cj16, 150 bp; Cj18, 207 bp; Cj131, 226 bp; CUJ131, 193 bp) that were not found in either C. rhombifer group or C. acutus. This individual was removed from subsequent STRUCTURE and NEWHYBRIDS analyses. A similar genetic treatment of C. palustris will be needed to accurately determine the actual paternity of this specimen. After the removal of nine admixed individuals, RC051 and MF271, we found that 20 alleles were C. rhombifer-a specific, 13 alleles were C. rhombifer-b specific, and 17 alleles were specific for C. acutus (Fig. 6). DISCUSSION Genetic status of C. rhombifer Biotic homogenization due to anthropogenic intervention can have serious evolutionary consequences on native species, such as changes in their global distribution (Olden et al., 2004). We have provided a genetic characterization of the Cuban crocodile using captive specimens, including some individuals with incongruent morphological and mitochondrial assignments. Phylogenetically, among New World crocodiles three separate clades were inferred from cyt-b sequence data (C. rhombifer-a and b, C. acutus– C. intermedius, and C. moreletii), with C. niloticus also grouping with the New World crocodiles (Fig. 2). Within C. rhombifer there were two distinct mitochondrial haplotype groups (a and b), but it is possible that greater haplotype diversity may be detected if a larger portion of the mitochondrial genome is sampled. These two haplotype groups were 0.9% divergent when comparing cyt-b sequences and 1.6% divergent when comparing D-loop sequences. If this pattern remains consistent, as the Caribbean is more extensively sampled for crocodiles, then C. rhombifer-b may represent a previously unidentified lineage. Sampling the nuclear genome (using microsatellites) we were able to detect both C. rhombifer-a specific alleles and C. rhombifer-b specific alleles, and thus two genetic clusters (C. rhombifer-I and C. rhombifer-II) were also inferred from model-based analyses of the microsatellite data. Taken together, these data may actually reflect the genetic diversity within the J. Exp. Zool. 8 WEAVER ET AL. Fig. 6. Allele frequency distribution for three genetic clusters inferred by model-based clustering methods (see Fig. 3), after inferred hybrids were removed. wild population, as the C. rhombifer species clade is consistently characterized as having two distinct genetic sub-groups. One of these exhibits ‘‘typical’’ C. rhombifer morphology whereas the other exhibits C. acutus morphology. We suspect that many crocodiles that have been ‘‘morphologically’’ identified as C. acutus in Cuba may actually belong to the C. rhombifer-b haplotype group. We can only speculate that b haplotypes may have been ancestrally present in Cuba, and that the current C. rhombifer-b group could be the result of past natural hybridization events. We cannot yet J. Exp. Zool. explain the evolutionary significance of C. rhombifer-b without reference samples from wild populations in Cuba, but action should be taken to identify these individuals and possibly remove them from captive breeding programs. Hybridization in crocodiles In Vietnam, captive C. rhombifer and C. siamensis have been deliberately hybridized (Thang, ’94). Hybridization was also reported to have occurred between captive C. rhombifer and GENETIC CHARACTERIZATION OF THE CUBAN CROCODILE C. acutus in breeding pens of the Laguna del Tesoro farm in Cuba (Ross, ’98), as well as in the wild (Ramos et al., ’94). Varona (’66) suggested that in Cuba at least some admixture was taking place between C. acutus and C. rhombifer, because several specimens exhibited morphological characters typical of both species, but clearly outside the normal range of variation found in C. acutus. However, among our samples we found that only MF271 and MF5267 exhibited anomalous morphology and that all F2 hybrids appeared to exhibit C. rhombifer morphology. We found that hybridization events involving captive C. rhombifer were invariably between the C. rhombifer-a group and C. acutus. All hybrid individuals had C. rhombiferlike mtDNA, which suggests that in captivity hybridization is typically between a female C. rhombifer and a male C. acutus. These results are consistent and congruent with those of FitzSimmons et al. (2002). Anthropogenic perturbation and natural migration events may pose a potential threat to the genetic integrity of C. acutus populations in the Caribbean, Florida, and Mesoamerica. Further, given the large proportion of admixture between C. acutus and C. moreletii (Cedeño-Vázquez et al., 2008; Rodriguez et al., 2008), populations of true crocodiles in Mexico may be threatened by an additional source of hybrid introgression from C. rhombifer. For example, RC051, which was found in waterways near Cancun, Mexico, carried a C. rhombifer-a haplotype and was morphologically identified as C. acutus; however, it was ultimately designated as an admixture between C. rhombifer-I and C. rhombifer-II by the nuclear data. Conservation and management implications Hybridization ultimately presents a management problem for New World crocodilians and complicates the identification of species based on morphology alone. A genetic evaluation (using both mtDNA and nDNA) in conjunction with a morphometric characterization can provide a more accurate view of an individual’s ancestry than either method alone. As an example, a crocodile marked for reintroduction into Cat Tien National Park (Vietnam) was genetically identified as a C. rhombifer  siamensis hybrid and was subsequently prevented from being released into the wild population (FitzSimmons et al., 2002). In order for reintroductions of native species to be 9 successful, only purebred individuals should be released back into their native habitats (Allendorf et al., 2001). Our work has provided an initial genetic assessment of the critically endangered Cuban crocodile. We hope these data can be used to identify pure individuals for breeding stock, which should be considered if repopulation of extirpated areas (e.g. Cienaga de Lanier) is to take place. An evaluation of current Cuban crocodile stocks is warranted to ensure a purebred captive breeding line, especially considering that out of seven US captive stocks surveyed five had some level of genetic admixture. Identification of hybrids using morphology may be problematic given that only two individuals studied were morphologically anomalous. Additionally, thorough genetic assessments of wild Cuban and American crocodile populations in the Caribbean are needed to provide a better genetic reference for assignment tests, to quantify the amount of potential genetic admixture between genetically differentiated groups, and to help clarify the evolutionary implications of the C. rhombifer-b sub-clade. The increase in ease and the concomitant decrease in cost of generating and analyzing molecular genetic data with statistical model-based analyses can be practical and informative for both management and conservation efforts. These methods are especially important for the conservation of endangered fauna with limited distributions, such as the Cuban crocodile. ACKNOWLEDGMENTS We thank R. Bradley and J. Hanson for assistance with sequencing, the TTU Core Lab for assistance with genotyping, and O. Sanjur and S. Mahecha at the Smithsonian Tropical Research Institute for laboratory assistance. We thank S. K. Davis and T. Guerra. In addition, we thank J. McVay for his input on data analyses, S. McCracken for curatorial assistance, M. A. Mullen, D. Hamilton, M. Vandewege, and P. Larsen for editorial assistance; and J. Isom, D. Fabing, and L. Durham for administrative assistance. This research was supported in part by a Howard Hughes Medical Institute grant through the Undergraduate Science Education Program to Texas Tech University. This research was developed under the following permits: SIM/A4-07, 2007/KY/000107, MX31441 (NRA: CFSTB2300411) and PRT 2-2996. J. Exp. Zool. 10 WEAVER ET AL. APPENDIX Samples used in this study, source of samples and group assignments for each data set. Sample # Morphology Source MtDNA1 LD042 LD068 LD069 LD099 LD100 LD123 LD125 LD145 LD147 LD148 LD150 LD154 LD178 LD179 LD182 LD183 LD197 LD198 LD221 LD305 LD324 MF267 MF268 MF269 MF270 MF271 MF272 MF274 MF277 MF1744 MF1745 MF5267 RC051 GC001 MF347 MF348 PM008 PM031 PM037 SP014 PM020 PM021 RC013 RC052 RC106 RC109 RC114 RC117 RC130 RC132 RC136 RC140 RC141 RC210 RC023 LD175 C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer Anomalous C. rhombifer C. rhombifer C. rhombifer C. rhombifer C. rhombifer Anomalous C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. acutus C. intermedius St. Augustine Zoo St. Augustine Zoo St. Augustine Zoo St. Augustine Zoo St. Augustine Zoo Gladys Porter Zoo Private property Private property Private property St. Augustine Zoo Jumbo Lair Jumbo Lair St. Augustine Zoo St. Augustine Zoo Toledo Zoo Toledo Zoo Bronx Zoo Bronx Zoo Bronx Zoo Private property Bronx Zoo St. Augustine Zoo St. Augustine Zoo Bronx Zoo Bronx Zoo Bronx Zoo Toledo Zoo Toledo Zoo Busch Gardens Gladys Porter Zoo Gladys Porter Zoo Jupiter, FL Cancun, Mexico Grand Cayman Island Jamaica Jamaica Busch Gardens Imperial River, FL Snapper Creek Canal, FL Private property North Key Largo, FL North Key Largo, FL Yucatan, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Quintana Roo, Mexico Yucatan, Mexico Quintana Roo, Mexico Private property a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a b b b b b b b – – – – – – – – – – – – – C. acutus C. acutus C. intermedius J. Exp. Zool. MtDNA2 a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a b b b b b b b C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. – acutus5 acutus5 acutus6 acutus6 acutus6 acutus6 acutus6 acutus6 acutus6 acutus6 acutus6 acutus6 acutus6 acutus6 acutus6 Msat3 Msat4 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 C.r.-1 Hyb C.r.-1 C.r.-1 Hyb Hyb Hyb Hyb Hyb C.r.-1 C.r.-1 Hyb Hyb C.r.-2 C.r.-2 C.r.-2 C.r.-2 C.r.-2 C.r.-2 C.r.-2 C.r.-2 C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. – – C.r.-I C.r.-I C.r.-I C.r.-I C.r.-I C.r.-I C.r.-I C.r.-I C.r.-I Hyb C.r.-I C.r.-I C.r.-I C.r.-I C.r.-I C.r.-I C.r.-I Hyb C.r.-I C.r.-I Hyb C.r.-I C.r.-I Hyb Hyb – Hyb Hyb C.r.-I C.r.-I Hyb Hyb C.r.-II C.r.-II C.r.-II C.r.-II C.r.-II C.r.-II C.r.-II C.r.-II C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. C.a. – – 11 GENETIC CHARACTERIZATION OF THE CUBAN CROCODILE Sample # RC206 RC207 Morphology C. moreletii C. moreletii Source Quintana Roo, Mexico Yucatan, Mexico MtDNA1 C. moreletii C. moreletii MtDNA2 C. moreletii C. moreletii Msat3 – – Msat4 – – Wild caught individuals. 1 Based cyt-b sequence data (see Fig. 2A). Based on tRNAPro- tRNAPhe-Dloop sequence data (see Fig. 2B). 3 Distance-based analysis of microsatellite data (see Fig. 3). 4 Model-based analysis of microsatellite data (see Fig. 4). 5 Rodriguez (2007). 6 Cedeño-Vázquez et al. (2008). 2 LITERATURE CITED Allendorf FW, Leary RF, Spruell P, Wenburg JK. 2001. The problems with hybrids: setting conservation guidelines. Trends Ecol Evol 16:613–622. Anderson EC, Thompson EA. 2002. A model-based method for identifying species hybrids using multilocus genetic data. Genetics 160:1217–1229. Bayliss P. 1987. Survey methods and monitoring within crocodile management programmes. In: Webb GJ, Manolis SC, Whithead PJ, editors. Wildlife management: crocodiles and alligators. Sydney, Australia: Surrey Beatty and Sons. p 157–175. Brochu CA. 2000. Phylogenetic relationships and divergence timing of Crocodylus based on morphology and the fossil record. Copeia 2000:657–673. Cavalli-Sforza LL, Edwards AWF. 1967. Phylogenetic analysis: models and estimation procedures. Am J Hum Genet 19:233–257. Cedeño-Vázquez JR, Rodriguez D, Calmé S, Ross JP, Densmore III LD, Thorbjarnarson JB. 2008. Hybridization between Crocodylus acutus and Crocodylus moreletii in the Yucatan Peninsula: I. Evidence from mitochondrial DNA and Morphology. J Exp Zool 309A. Chavananikul V, Wattanodorn S, Youngprapakorn P. 1994. Karyotypes of 5 species of crocodile kept in Samutprakan crocodile farm and zoo. In: Crocodiles, Proceedings of the 12th Working Meeting of the Crocodile Specialist Group, IUCN—The World Conservation Union, Gland, Switzerland. p 58–62. Cohen MM, Gans C. 1970. The chromosomes of the order crocodilia. Cytogenetics 9:81–105. Densmore LD. 1983. Biochemical and immunological systematics of the order crocodilia. In: Hecht MK, Wallace B, Prance GT, editors. Evolutionary biology. Vol. 16. New York: Plenum Publ Co. p 397–465. Dever JA, Densmore III LD 2001. Microsatellites in Morelet’s crocodile (Crocodylus moreletii) and their utility in addressing crocodilian population genetics questions. J Herpetol 35:541–544. FitzSimmons NN, Tanksley S, Forstner MRJ, Louis EE, Daglish R, Gratten J, Davis S. 2001. Microsatellite markers for Crocodylus: new genetic tools for population genetics, mating system studies and forensics. In: Grigg GC, Seebacher F, Franklin CE, editors. Crocodilian biology and evolution. Chipping Norton: Surrey Beatty & Sons. FitzSimmons NN, Buchan JC, Lam PV, Polet G, Hung TT, Thang NQ, Gratten J. 2002. Identification of purebred Crocodylus siamensis for reintroduction in Vietnam. J Exp Zool 294:373–381. Frankham R, Ballou JD, Briscoe DA. 2002. Introduction to conservation genetics. Cambridge: Cambridge University Press. Franz R, Morgan GS, Albury N, Buckner SD. 1995. Fossil skeleton of a cuban crocodile (Crocodylus rhombifer) from a blue hole on Abaco Bahamas. Caribbean J Sci 31:149–152. Glenn TC, Staton JL, Vu AT, Davis LM, Bremer JRA, Rhodes WE, Brisbin Jr. IL, Sawyer RH. 2002. Low mitochondrial DNA variation among American alligators and a novel non-coding region in crocodilians. J Exp Zool 294:312–324. Gorzula S, Arocha-Pinango CL, Salazar C. 1976. A method of obtaining blood by vein puncture from large reptiles. Copeia 1976:838–839. Hall TA. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/ NT. Nucleic Acids Symp Ser 41:95–98. Hekkala ER. 2004. Conservation genetics at the species boundary: case studies from African and Caribbean crocodiles (Genus: Crocodylus). PhD Dissertation. Columbia University, New York, NY. Huelsenbeck JP, Ronquist FR. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754–755. Langella. 1999. Populations v1.2.28: Available at http:// www.cnrs-gif.fr/pge/bioinfo/populations. Marshall TC, Slate J, Kruuk LEB, Pemberton JM. 1998. Statistical confidence for likelihood-based paternity inference in natural populations. Mol Ecol 7:639–655. Morgan GS, Richard F, Crombie RI. 1993. The Cuban crocodile, Crocodylus rhombifer, from late quaternary fossil deposits on Grand Cayman. Caribbean J Sci 29:153–156. Nylander JAA. 2004. MrModeltest 2.1: Program distributed by the author. Evolutionary Biology Centre, Uppsala. Olden JD, Poff NL, Douglas MR, Douglas ME, Fausch KD. 2004. Ecological and evolutionary consequences of biotic homogenization. Trends Ecol Evol 19:18–24. Page RDM. 1996. TreeView: an application to display phylogenetic trees on personal computers. Comput Appl Biosci 12:357–358. Posada D, Crandall KA. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817–818. Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945–959. Ramos R, de Buffrenil V, Ross JP. 1994. Current status of the Cuban crocodile, Crocodylus rhombifer, in the wild. In: Crocodiles, Proceedings of the 12th Working Meeting of the J. Exp. Zool. 12 WEAVER ET AL. Crocodile Specialist Group, IUCN—The World Conservation Union, Gland, Switzerland. p 113–140. Ray DA, Densmore L. 2002. The crocodilian mitochondrial control region: general structure, conserved sequences, and evolutionary implications. J Exp Zool 294:334–345. Ray DA, Dever JA, Platt SG, Rainwater TR, Finger AG, McMurry ST, Batzer MA, Barr B, Stafford PJ, McKnight J, Densmore LD. 2004. Low levels of nucleotide diversity in Crocodylus moreletii and evidence of hybridization with C. acutus. Conserv Genet 5: 449–462. Rodriguez D. 2007. Crocodilian evolution, systematics and population genetics: recovery and ecological interactions of the American crocodile (Crocodylus acutus). PhD Dissertation. Texas Tech University, Lubbock, Texas. Rodriguez D, Cedeño-Vázquez JR, Forstner MRJ, Densmore III LD. 2008. Hybridization between Crocodylus acutus and Crocodylus moreletii in the Yucatan Peninsula: II. Evidence from microsatellites. J Exp Zool 309A. Ross JP. 1998. Crocodiles status survey and conservation action plan. Gainesville, FL: Crocodile Specialist Group. 96p. J. Exp. Zool. Swofford DL. 2002. PAUP v4.0b10: phylogenetic analysis using parsimony ( and other methods). Version 4.0b10. Sunderland, MA: Sinauer. Thang NQ. 1994. The status of Crocodylus rhombifer in the Socialist Republic of Vietnam. In: Crocodiles, Proceedings of the 12th Working Meeting of the Crocodile Specialist Group, IUCN—World Conservation Union. Gland, Switzerland. p 141–151. Thompson J, Gibson T, Plewniak F, Jeanmougin F, Higgins D. 1997. The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 25:4876–4882. Thorbjarnarson J, Mazzotti F, Sanderson E, Buitrago F, Lazcano M, Minkowski K, Muniz M, Ponce P, Sigler L, Soberon R, Trelancia AM, Velasco A. 2006. Regional habitat conservation priorities for the American crocodile. Biol Conserv 128:25–36. Varona LS. 1966. Notas sobre los crocodı́lidos de Cuba y descripción de una nueva especie del Pleistoceno. Poeyana 16:1–21. Varona LS. 1984. Los cocodrilos fosiles De Cuba (reptilia: crocodylidae). Caribbean J Sci 20:13–18.