Volume 55, Issue 2 pp. 167-174
Original Article
Full Access

The leaf turtle population of Phnom Kulen National Park (northwestern Cambodia) has genetic and morphological signatures of hybridization

Melita Vamberger

Corresponding Author

Melita Vamberger

Museum of Zoology, Senckenberg, Dresden, Germany

Corresponding author: Melita Vamberger ([email protected])Search for more papers by this author
Louise Durkin

Louise Durkin

Department of Ecology, Environment and Evolution, La Trobe University, Bundoora, VIC, Australia

Search for more papers by this author
Chamnan Kim

Chamnan Kim

Centre for Biodiversity Conservation, Department of Biology, Royal University of Phnom Penh, Phnom Penh, Cambodia

Search for more papers by this author
Markus Handschuh

Markus Handschuh

Angkor Centre for Conservation of Biodiversity, Phnom Kulen National Park, Siem Reap Province, Cambodia

Search for more papers by this author
Rattanak Seng

Rattanak Seng

General Department for Administration of Nature Conservation and Protection, Ministry of Environment, Phnom Penh, Cambodia

Search for more papers by this author
Uwe Fritz

Uwe Fritz

Museum of Zoology, Senckenberg, Dresden, Germany

Search for more papers by this author
First published: 03 February 2017
Citations: 6
Contributing authors: Louise Durkin ([email protected]), Chamnan Kim ([email protected]), Markus Handschuh ([email protected]), Rattanak Seng ([email protected]), Uwe Fritz ([email protected])

Abstract

Cambodia is known to harbour three distinct species of Southeast Asian leaf turtles (Cyclemys spp.), which are heavily traded and common in seizures of wildlife. Confiscated leaf turtles are often released to natural habitats. Thus, an exact knowledge of the distribution of the individual species is of great importance to avoid the introduction of non-native turtles, posing the risk of competition and hybridization. In this study, we examine a recently discovered leaf turtle population from Phnom Kulen National Park using external morphology, 17 unlinked microsatellite loci and the mitochondrial cytochrome b gene. Leaf turtles from the Phnom Kulen National Park morphologically resemble C. oldhamii, but harbour mitochondrial haplotypes of C. atripons. With respect to microsatellite loci, the turtles are distinct from C. atripons. Unfortunately no material of C. oldhamii was available from Cambodia. We propose that the Phnom Kulen population represents either a natural hybrid swarm of C. atripons and C. oldhamii or a distinct undescribed species with introgressed mitochondria of C. atripons. This underlines that genetic differentiation of wild leaf turtle populations in Cambodia is complex and suggests that this differentiation pattern becomes increasingly threatened by translocations of confiscated individuals. For drawing a definite conclusion about the taxonomic status of the Phnom Kulen population, denser sampling of other Cambodian leaf turtle populations would be required, in particular of C. oldhamii.

Introduction

Until recently, the Southeast Asian leaf turtle genus Cyclemys (Testudines: Geoemydidae) was thought to contain only one or two species, Cyclemys dentata (Gray, 1831) and C. tcheponensis (Bourret, 1939). However, starting in the mid-1990s, several studies resurrected and described additional leaf turtle species (Fritz et al. 1996, 1997, 2008; Iverson and McCord 1997; Stuart and Fritz 2008), so that currently seven distinct species are recognized (van Dijk et al. 2014). Cambodia is thought to harbour three species: C. atripons Iverson and McCord, 1997; C. oldhamii Gray, 1863 and C. pulchristriata Fritz et al., 1997; although the exact distribution ranges are not well understood (van Dijk et al. 2014). Yellow-bellied leaf turtles from southwestern Cambodia represent C. atripons, while yellow-bellied leaf turtles from the southeast of the country belong to C. pulchristriata, a species found mainly in Vietnam. Dark-bellied leaf turtles from the inland regions of Cambodia are assigned to C. oldhamii (Iverson and McCord 1997; Fritz et al. 2008; Stuart and Fritz 2008; van Dijk et al. 2014).

Leaf turtles are among the more frequently encountered turtle species in Cambodia, as indicated both by wildlife surveys and confiscations from the illegal wildlife trade (Som et al. 2006; Som and Kheng 2007; TRAFFIC 2014). They can be found in various freshwater habitats, including rivers and streams, swamps, grassland ponds and flooded rice paddies. Adult leaf turtles also utilize terrestrial habitats nearby freshwater (Durkin 2012). Leaf turtles are heavily traded and common in seizures of smuggled wildlife en route from Cambodia to food and traditional medicine markets in Vietnam, indicating that wild populations continue to be depleted. In 2013, all Cyclemys species were placed onto CITES Appendix II (CITES 2013). In Cambodia, confiscated turtles are often released (pers. observ.). To avoid introduction of non-native turtles with the risks of competition and hybridization, the exact knowledge of the local distribution of the different leaf turtle species is of great importance.

Durkin et al. (2010) reported dark-bellied leaf turtles from Phnom Kulen National Park in northwestern Cambodia which they tentatively identified as C. aff. atripons. However, the turtles' morphology conflicted with this determination, and based on coloration and pattern, Kim (2011) and Durkin (2012) assigned the Phnom Kulen population to C. oldhamii. Phnom Kulen National Park is located in between the known distribution ranges of C. atripons and C. oldhamii and is geographically isolated in terms of Cyclemys records. Thus, the taxonomic identity of the Phnom Kulen population remains unclear. In this study, we compare Cyclemys samples from Phnom Kulen National Park with previously published mitochondrial DNA (mtDNA) sequence data of all currently recognized Cyclemys species. In addition, we compare the Phnom Kulen turtles with representatives of C. atripons by genotyping them at 17 unlinked microsatellite loci and compare their coloration and pattern with C. atripons and C. oldhamii.

Materials and Methods

Sampling and laboratory procedures

Samples from 62 leaf turtles were used for analysis (Appendix). Twelve of these were Cyclemys atripons, most of which had been previously studied (Fritz et al. 2008). These turtles conform morphologically to the original species description (Iverson and McCord 1997). The remaining 50 individuals were sampled between 2010 and 2012 at Kbal Spean in western Phnom Kulen National Park, northwestern Cambodia. The turtles were captured using baited funnel traps set in the Kbal Spean River (Durkin et al. 2010; Kim 2011; Durkin 2012). Samples were either tail tips or marginal scute clippings stored in 95% ethanol. DNA was extracted using the innuPREP DNA Mini Kit (Analytik Jena AG, Jena, Germany) or the NucleoSpin Tissue Kit (Macherey-Nagel, Düren, Germany).

As a mitochondrial marker, the cytochrome b gene (cyt b) was sequenced for 45 freshly collected samples (Appendix) using the PCR primers CytbG (Spinks et al. 2004) and mt-E-Rev2 (Fritz et al. 2006) or mt-c-For2 and mt-f-na (Fritz et al. 2006). PCR was performed in a final volume of 20 μl using 1 unit DFS-Taq polymerase (Bioron GmbH, Ludwigshafen, Germany) with the buffer recommended by the supplier (2 μl PCR buffer 10× incl. MgCl2) and a final concentration of 0.2 mM of each dNTP (Thermo Fisher Scientific Biosciences GmbH, St. Leon-Rot, Germany), 0.5 μM of each primer (biomers.net GmbH, Ulm, Germany) and approximately 10–40 ng of total DNA. For PCR, an initial 5-min-long denaturation step at 95°C was followed by 35–40 cycles with denaturation at 95°C for 30 s, annealing at 55°C for 30 s, extension at 72°C for 1 min and a final elongation step of 10 min at 72°C. PCR products were purified using the ExoSAP-IT enzymatic clean-up (Affymetrix USB, Cleveland, OH, USA; modified protocol: 0.2 μl ExoSAP-IT per sample 30 min at 37°C, 15 min at 80°C) and sequenced using the forward primer mt-c-For2 and the reverse primer mt-E-Rev2 on an ABI 3730 Genetic Analyzer (Life Technologies, Carlsbad, CA, USA) using the BigDye Terminator v3.1 Cycle Sequencing Kit (Life Technologies). Cycle sequencing reactions were purified by gel filtration using the Performa DTR V3 96-Well Short Plate Kit (EdgeBio, Gaithersburg, MD, USA) and 400 μl of a 5% Sephadex solution (GE Healthcare, München, Germany). DNA sequences were aligned using bioedit 7.0.5.3 (Hall 1999) and trimmed to a length of 1166 bp. This fragment comprised the complete cyt b gene (1140 bp) plus adjacent DNA coding for tRNA-Thr (26 bp).

In addition, 17 unlinked microsatellite loci (Table S1) were analysed for each sample. Two to four loci each were combined in multiplex PCRs; one locus (GP81) was processed alone (Table S1). The final volume of each PCR was 10 μl, containing 0.5 unit Taq polymerase (Bioron, Ludwigshafen, Germany) with the incomplete buffer recommended by the supplier and a final concentration of 1.5 mM MgCl2 (Bioron), 0.2 mM of each dNTP (Fermentas, St. Leon-Rot, Germany), 2.5 μg bovine serum albumin (Fermentas), approximately 10–20 ng of total DNA, 0.25 μM forward and reverse primers. Forward primers were fluorescent-labelled (Table S1). PCR cycling conditions were: After 5-min denaturation at 94°C, 35 cycles were run with denaturation at the same temperature for 45 s, annealing as in Table S1 for 1 min and elongation at 72°C for 1 min, followed by a 10-min-long final elongation step. PCR products were diluted with water in a ratio of 1:100. Fragment lengths were determined using an ABI 3730 Genetic Analyzer, the GeneScan-600 LIZ Size Standard and the software genemapper (all: Life Technologies).

Data analysis

mtDNA

Phylogenetic relationships of previously published mtDNA sequences (Fritz et al. 2008) combined with the newly generated data from this study (Appendix) were inferred by Maximum Likelihood (ML) analyses using RAxML 7.2.6 (Stamatakis 2006), the implemented evolutionary GTR + G model, and a GenBank sequence of Heosemys spinosa (AY434578) as outgroup. Five independent ML searches were performed with different starting conditions and the fast bootstrap algorithm to explore the robustness of the results by comparing the best trees. Subsequently, 1000 non-parametric thorough bootstrap replicates were calculated and the values plotted against the best tree.

Using the aligned sequence data, the mutational relationships of haplotypes were also displayed as a parsimony network with TCS 1.21 (Clement et al. 2000).

Microsatellites

For the Phnom Kulen turtles and the Cyclemys atripons studied, pairwise linkage disequilibrium between loci and Hardy–Weinberg equilibrium were tested using arlequin 3.5.2.1 (Excoffier and Lischer 2010). Furthermore, the 17 microsatellite loci were analysed using an unsupervised Bayesian clustering approach as implemented in structure 2.3.4 (Pritchard et al. 2000; Hubisz et al. 2009). In doing so, the admixture model and correlated allele frequencies were used. structure searches in the data set for partitions which are, as far as possible, in Hardy–Weinberg equilibrium and linkage equilibrium. Unsupervised analyses were chosen because this approach clusters samples strictly according to their genetic information, but without any presumptions about population structuring (e.g. geographical distances, sampling sites). Because microchecker 2.2.3 (van Oosterhout et al. 2004) suggested the presence of null alleles for three loci (Mauca01, Maucas14, Test56) in the Phnom Kulen population, the data set was corrected for null alleles according to Falush et al. (2007). All calculations were run for = 1–10, and the most likely number of clusters (K) was determined using the ∆K method (Evanno et al. 2005) with structure harvester (Earl and vonHoldt 2012). Calculations were repeated 10 times for each K using a MCMC chain of 750 000 generations for each run, including a burn-in of 250 000 generations. Population structuring and individual admixture were then visualized using distruct 1.1 (Rosenberg 2004).

Diversity within the Phnom Kulen population and comparison with Cyclemys atripons

For the Phnom Kulen population and Cyclemys atripons, diversity and divergence parameters were estimated using microsatellite data. For doing so, number and size of microsatellite alleles were compared by a frequency table produced in convert 1.31 (Glaubitz 2004). For inferring locus-specific observed (HO) and expected heterozygosities (HE) and for performing a locus-by-locus analysis of molecular variance (AMOVA; 10 000 permutations), arlequin 3.5.2.1 was used. Locus-specific excess or deficiency of heterozygotes as expressed by the inbreeding coefficient FIS (Weir and Cockerham 1984) was calculated with fstat 2.9.3.2 (Goudet 1995). The same software was also used for computing values for locus-specific allelic richness and testing statistical significance of FIS for each locus and across all loci using randomizations and Bonferroni correction (Rice 1989).

Coloration and pattern

From each turtle from Phnom Kulen National Park, morphometric characters were taken to the nearest millimetre using callipers. The animals were weighed to the nearest gram using spring scales or electronic scales and uniquely marked using either a microchip, a marginal scute notch or both. Each turtle was photographed in dorsal and ventral aspect. In addition, photographs of the temporal region were taken, where possible. Recapture of marked turtles over several years allowed observations of the progression of plastral patterning, particularly in younger individuals that show more growth and change from year to year.

Results

Mitochondrial relationships

All individuals of the Phnom Kulen population yielded the same mitochondrial haplotype (k1; Fig. 1; Appendix) which differed in parsimony network analysis by seven mutational steps from the haplotypes of Cyclemys atripons. However, among the haplotypes of C. atripons, up to 12 mutations were observed (Fig. 1). Accordingly, in ML analyses, the haplotype of the Phnom Kulen population clustered with maximum support with the sequences of Catripons and within this clade was sister to the remaining haplotypes (Fig. S1).

Details are in the caption following the image
Genotypic structuring of 60 leaf turtles for = 2 using 17 microsatellite loci (top). Shown is the structure run with the best probability value. Within each cluster, an individual is represented by a vertical column that reflects its ancestry; individuals with mixed ancestries would show differently coloured segments. The mitochondrial identity (Appendix) of each sample is indicated below the structure diagram (red = haplotypes of Cyclemys atripons, white = missing data). Bottom left: Geographical distribution of studied leaf turtles; for turtles with mitochondrial sequences, the haplotype names are given (C. atripons, a1-a6, n = 8; leaf turtles from Phnom Kulen National Park, k1, n = 37). Sampling sites are colour-coded according to the structure clusters. Bottom right: Parsimony network showing mutational relationships of haplotypes. Symbol size corresponds to haplotype frequency. Lines connecting haplotypes represent one mutational step, if not otherwise indicated.

Population genetic analyses using microsatellite data

The ∆K method of Evanno et al. (2005) revealed = 2 as the optimal number of structure clusters (Fig. S2), with one cluster corresponding to leaf turtles from the Phnom Kulen population and the other to Cyclemys atripons, without any indication of admixture (Fig. 1). Across the two clusters, the numbers of alleles per locus ranged from 1 to 22; from a total of 165 alleles, 114 private ones were found. Genetic diversity indices of the two clusters were similar, despite the difference in the sample size. In the Phnom Kulen population, pronounced linkage disequilibrium (6%) was evident (Tables 1 and 2). The Phnom Kulen population and Catripons differed by an FST value of 0.21; thus, 21% of the observed global molecular variance occurred among and 79% within clusters.

Table 1. Genetic diversity of the 17 microsatellite loci in Phnom Kulen population and Cyclemys atripons
Locus n A AR H O H E F IS HWE Null alleles
PKNP atripons PKNP atripons PKNP atripons PKNP atripons PKNP atripons PKNP atripons PKNP atripons
Maucas01 6 7 4.000 5.539 0.34694 0.30000 0.49548 0.82632 0.302* 0.649* No No Yes Yes
Maucas06 4 9 7.918 3.679 0.73469 0.80000 0.67873 0.60526 −0.083 −0.346 Yes Yes No No
Maucas12 2 4 2.940 2.000 0.50000 0.50000 0.40343 0.39474 −0.242 −0.286 Yes Yes No No
Maucas14 9 12 8.000 7.414 0.58000 0.80000 0.80081 0.87895 0.278* 0.094 No Yes Yes No
Maucas17 5 5 2.000 4.618 0.22000 0.90000 0.28505 0.77368 0.230 −0.174 Yes Yes No No
Maucas18 6 11 8.937 5.865 0.90000 0.90000 0.82768 0.86316 −0.088 −0.045 Yes Yes No No
Test56 5 6 4.000 4.660 0.22000 0.11111 0.55697 0.62092 0.607* 0.830* No No Yes Yes
Maucas20 5 5 2.000 4.603 0.58000 0.70000 0.47091 0.75263 −0.235 0.074 Yes Yes No No
Emys2 10 16 10.000 8.100 0.98000 0.90000 0.88646 0.89474 −0.107 −0.006 Yes Yes No No
MR3 5 13 8.937 5.000 0.80000 0.28571 0.73859 0.82418 −0.084 0.671* Yes No No Yes
GP19 1 4 3.937 1.000 0.42000 n/a 0.44949 n/a 0.066 n/a No Yes No No
GmuD16 5 9 7.000 4.333 0.79167 0.44444 0.79386 0.66667 0.003 0.347 Yes Yes No No
GmuB08 5 6 5.000 4.511 0.80000 0.33333 0.72707 0.61438 −0.101 0.473* Yes No No Yes
Test10 11 22 17.835 9.391 0.93878 1 0.91269 0.93464 −0.029 −0.075 Yes Yes No No
TWL221 3 4 2.997 2.778 0.20000 0.55556 0.18505 0.54248 −0.082 −0.026 Yes Yes No No
TWI61 6 20 16.814 5.072 0.94000 0.55556 0.91273 0.56209 −0.030 0.012 Yes Yes No No
GP81 9 12 6.000 7.954 0.72340 0.88889 0.70991 0.87582 −0.019 −0.016 Yes Yes No No
  • PKNP, Phnom Kulen National Park population; atripons, Cyclemys atripons; nA, number of alleles; AR, allelic richness; HO, average observed heterozygosity; HE, average expected heterozygosity; FIS, inbreeding coefficient; HWE, Hardy–Weinberg equilibrium. FIS values indicated by an asterisk are statistically significant. Bold script highlights absence of HWE or presence of null alleles, respectively.
Table 2. Genetic diversity in the Phnom Kulen population and Cyclemys atripons
n n A n Ā n p AR H O H E F IS LD
PKNP 50 97 5.71 68 6.960 0.63 0.64 0.02 6.00
C. atripons 10 165 9.71 46 5.089 0.59 0.68 0.13* 0.72
  • PKNP, Phnom Kulen National Park population; n, number of individuals; nA, number of alleles; nĀ, average number of alleles; nP, number of private alleles; AR, allelic richness; HO, average observed heterozygosity; HE, average expected heterozygosity; FIS, average inbreeding coefficient; LD, linkage disequilibria, per cent of pairwise comparisons between microsatellite loci. The FIS value with an asterisk is statistically significant.

Coloration and pattern

The leaf turtles from Phnom Kulen National Park had straight carapace lengths from 5.8 to 21.0 cm. All corresponded in coloration and pattern perfectly to the eastern coloration morph of Cyclemys oldhamii as described in Fritz et al. (2008). Phnom Kulen leaf turtles had a speckled pattern on the crown of the head and salmon striping on the neck on a dark background. The throat typically showed a mottled dark pattern on salmon background. The carapace of Phnom Kulen turtles was generally dark brownish. The plastron showed in the smallest individuals a spotted pattern on yellow background. With increasing size and age, this pattern was replaced by densely radiating dark streaks on each plastral scute. Aged individuals showed a gradient of patterns from dense dark brown radiations on a yellow background to a uniformly dark coloured plastron, as typical for C. oldhamii (Fig. 2).

Details are in the caption following the image
Ventral aspects of leaf turtles from Cambodia. Not to scale. Phnom Kulen National Park, northwestern Cambodia: (a) small juvenile, 5.8 cm straight carapace length; (b) medium-sized juvenile with radiating dark pattern, 12.3 cm; (c) adult with fading radiating pattern and nearly uniform dark plastron, 19.8 cm. Phnom Samkos Wildlife Sanctuary, southwestern Cambodia: (d) Cyclemys atripons, adult, 17.8 cm straight carapace length, (Centre for Biodiversity Conservation, Royal University of Phnom Penh, CBC 01515).

Discussion

The taxonomy and species number of Southeast Asian leaf turtles was recently much in flux, with several species described as new to science or resurrected from synonymy (Fritz et al. 1996, 1997, 2008; Iverson and McCord 1997). The leaf turtles of Phnom Kulen National Park morphologically closely resemble Cyclemys oldhamii, a dark-bellied leaf turtle species which has been recorded at a distance of approximately 190 km from the national park in the Prey Long forest of central Cambodia (Som and Kheng 2007). Yet, in phylogenetic analyses the mitochondrial haplotypes of the Phnom Kulen turtles do not cluster with C. oldhamii (Fig. S1), but with C. atripons, another leaf turtle species that is morphologically clearly distinct and known to occur in the Cardamom Mountains of southwest Cambodia (Stuart and Platt 2004; Som et al. 2006). The nearest unambiguous record of C. atripons to Phnom Kulen National Park are two specimens collected approximately 175 km away in Phnom Samkos Wildlife Sanctuary in the Cardamom Mountains and deposited at the Centre for Biodiversity Conservation in Phnom Penh under collection numbers CBC 00576 and CBC 01515. However, a questionable extraterritorial record for C. atripons has recently been published from Krong Samraong District in northwestern Cambodia (Brakels et al. 2016), approximately 90 km from the park.

Cyclemys atripons belongs to a phylogenetically well-supported clade comprised of three yellow-bellied species, C. atripons, C. dentata and C. pulchristriata. Cyclemys atripons is the sister taxon to C. dentata, a species distributed across the Great Sunda Islands and the Malay Peninsula (Fritz et al. 2008). The leaf turtles from Phnom Kulen National Park harbour mitochondrial haplotypes of C. atripons, but our analyses of microsatellite DNA provide evidence that they are genetically distinct from the latter species. Thus, we can dismiss the hypothesis that the Phnom Kulen turtles belong to the species C. atripons and that this species is polymorphic. Considering that the Phnom Kulen turtles are morphologically indistinguishable from C. oldhamii, it seems reasonable to hypothesize that the Phnom Kulen population represents a hybrid swarm of C. atripons and C. oldhamii. This hypothesis is also supported by the pronounced linkage disequilibrium found in the Phnom Kulen turtles and by the lack of Hardy–Weinberg equilibrium at several loci (Tables 1 and 2), as typical for hybridized populations. That also some loci of C. atripons are not in Hardy–Weinberg equilibrium is not astonishing because our sample of this species is very small and does not represent a natural population (unlike the Phnom Kulen turtles), but an assemblage of individuals from several distinct populations.

The illegal wildlife trade in Southeast Asia is extensive, and confiscations over the last two decades have included many hundreds of kilograms of live turtles and tortoises being smuggled from Cambodia to Vietnam and China for the food and traditional medicine markets (Compton 2000; own unpubl. observ.). Seizures of live animals are often released back into the wild without knowledge of their origin. For example, in 2008, 419 kg of live pythons and turtles of uncertain origin confiscated in Battambang, Cambodia, were released to the wild, including the Tonle Sap lake (TRAFFIC 2014). This practice has obvious implications for the genetic integrity of species and can threaten the persistence of morphologically or genetically distinct populations through gene pool mixing (Laikre et al. 2010). In certain conservation contexts, a low level of assisted immigration can produce fitness benefits for populations with outbreeding depression (‘genetic rescue’; Tallmon et al. 2004), but for species that do not require immediate conservation intervention, gene pool mixing can erode evolutionarily significant units (ESUs) and complicate baseline taxonomic work. While to our knowledge no bulk releases of leaf turtles have occurred in Phnom Kulen National Park, the study site Kbal Spean is a culturally significant site, with a history of ceremonial human presence dating back over 1000 years (Freeman et al. 1999). In Cambodia, the tradition of releasing captured or purchased turtles for religious reasons persists today (Ihlow et al. 2016). Whether releases of turtles occurred at Kbal Spean over the centuries due to human intervention is unknown. However, it is unlikely that the genetic situation found in the Phnom Kulen population is caused by the translocation of a few alien leaf turtles because the entire population seems affected. All studied leaf turtles were genetically similar in that they harboured the same mitochondrial haplotype and were assigned to the same cluster in structure analyses. In addition, both the Phnom Kulen population and C. atripons possess many private alleles (Table 2). This would not be expected if the Phnom Kulen population is derived from recent hybridization because in the latter case, a high number of shared alleles with C. atripons would be expected.

Unfortunately, no samples of Cambodian C. oldhamii were available for genetic examination, and our conclusion that the Phnom Kulen turtles are a hybrid swarm remains somewhat speculative. In particular, the inferred genetic impact of C. oldhamii is backed only by morphology. However, morphology alone is very convincing in this case and introgressed mitochondria have been previously described for another pair of yellow-bellied and dark-bellied Cyclemys species, C. dentata and C. enigmatica. The yellow-bellied C. dentata occurs throughout its distribution range sympatrically with a dark-bellied leaf turtle of which only few individuals have been characterized genetically (Fritz et al. 2008). All of these turtles yielded mitochondrial haplotypes of C. dentata, while nuclear genomic markers were distinct, which is why Fritz et al. (2008) described these dark-bellied turtles as the distinct species C. enigmatica. Thus, we also cannot exclude the possibility that the Phnom Kulen turtles represent an undescribed species with introgressed mitochondria from C. atripons. However, for the time being we refrain from taxonomic consequences. Further research, including extensive sampling of other leaf turtle populations from Cambodia, especially of C. oldhamii, is needed to clarify this situation.

Our study highlights that the taxonomy of leaf turtles is still incompletely known and that additional investigations will contribute to a significantly better understanding of their taxonomic and genetic diversity. Even though our results do not support the hypothesis that the observed genetic constitution of the leaf turtle population in Phnom Kulen National Park has been noticeably impacted by the introduction of leaf turtles from elsewhere, we provide evidence for a complicated genetic variation in wild populations that may become increasingly threatened by translocations of confiscated individuals.

Acknowledgements

We are grateful to H. E. Chay Smith (General Director of the General Department for Administration of Nature Conservation and Protection GDANCP, Ministry of Environment of the Royal Government of Cambodia) for kindly issuing the relevant permits. We thank Sy Ramony (Director of Department of National Park and Wildlife Sanctuary Director of Department of National Park and Wildlife Sanctuary of GDANCP, Ministry of Environment) and Hong Daravuth (Deputy Director of the Department of National Park and Wildlife Deputy Director of the Department of National Park and Wildlife Sanctuary of GDANCP and Director of Protected Areas in Siem Reap Province, Ministry of Environment) for their generous support. Field work in Phnom Kulen National Park was partially funded by Brian Malone (Department of Ecology, Environment and Evolution, La Trobe University) and the Angkor Centre for Conservation of Biodiversity (ACCB). David Emmett (Conservation International) provided samples of Cyclemys atripons. Thanks for lab work go to Anja Rauh (Senckenberg Dresden).

    Appendix:

    Studied leaf turtle samples, geographical sampling localities (WGS84), mitochondrial haplotypes (mtDNA, see Fig. 1) and cluster assignment in structure analyses. All leaf turtles from Phnom Kulen National Park yielded the same mitochondrial haplotype (ENA accession number LT595721); the newly sequenced sample of Cyclemys atripons (8588) had the same haplotype as a previously published turtle (AM931626). For accession numbers of other previously published mtDNA sequences, see Fritz et al. (2008).

    Sample Locality mtDNA = 2 X Y
    17 Unknown n/a Red n/a n/a
    1382 Cambodia: Koh Kong: Sre Ambel District: Sre Ambel a6 Red 11.1222 103.7458
    1383 Cambodia: Koh Kong: Sre Ambel District: Sre Ambel a5 n/a 11.1222 103.7458
    4487 Cambodia: Koh Kong: Tatai River, Central Cardamoms Protection Forest a4 Red 11.8214058 103.53540227
    4488 Cambodia: Koh Kong: upper Tatai River, Central Cardamoms Protection Forest n/a Red 11.8214058 103.53540227
    4489 Cambodia: Koh Kong: camp on Areng River, Central Cardamoms Protection Forest n/a Red 11.6373879 103.5730672
    4490 Cambodia: Koh Kong: mid-reaches of the Tatai River, Central Cardamoms Protection Forest a2 Red 11.6878695 103.6110616
    4491 Cambodia: Koh Kong: coastal forests, southern Cardamoms, near Trapeang Rung, estuary a1 Red 11.3775173 103.2586019
    4492 Cambodia: Koh Kong: coastal forests, southern Cardamoms, near Trapeang Rung, estuary a1 Red 11.3775173 103.2586019
    4493 Cambodia: Koh Kong: Tatai Krom village, flooded forest near river estuary, southern Cardamoms n/a Red 11.5578674 103.1414840
    8588 Cambodia: Koh Kong: Phnom Samkos Wildlife Sanctuary, northern Cardamom Mountains a1 Red 12.20419 103.07009
    42 516 Unknown a3 n/a n/a n/a
    6412 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 yellow 13.68436609 104.0169064
    6413 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.68175921 104.020699
    6415 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.68660597 104.0186077
    6416 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.6949395 104.0112127
    8548 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.68675 104.01539
    8549 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.68723 104.01516
    8550 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.68723 104.01516
    8551 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.68948 104.0147
    8553 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69089 104.01386
    8556 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.6924 104.01292
    8557 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 yellow 13.6924 104.01292
    8559 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69524 104.01055
    8560 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69524 104.01055
    8561 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 yellow 13.69547 104.01015
    8562 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69547 104.01015
    8563 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69751 104.00636
    8564 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69808 104.00642
    8565 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69774 104.00498
    8566 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.68442 104.01686
    8567 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69177 104.01341
    8568 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69487 104.01079
    8569 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69524 104.01055
    8570 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69524 104.01055
    8571 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69548 104.0096
    8572 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69548 104.0096
    8573 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69774 104.00498
    8574 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.68992 104.01448
    8575 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69177 104.01341
    8576 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.68948 104.0147
    8577 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69548 104.0096
    8578 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69789 104.00679
    8579 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69295 104.01207
    8580 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69049 104.01389
    8581 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69049 104.01389
    8582 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69049 104.01389
    8584 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69774 104.00498
    8586 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69774 104.00498
    8587 Cambodia: Kbal Spean River, Phnom Kulen National Park k1 Yellow 13.69089 104.01386
    10 489 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.68914048 104.0148154
    10 491 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69497418 104.0108427
    10 492 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69734267 104.0062649
    10 493 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69362979 104.0115234
    10 494 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69362979 104.0115234
    10 495 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69542353 104.0101658
    10 497 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69107122 104.0137995
    10 498 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69154039 104.0135571
    10 499 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69734267 104.0062649
    10 500 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.69542353 104.0101658
    10 501 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.68950153 104.0146659
    10 502 Cambodia: Kbal Spean River, Phnom Kulen National Park n/a Yellow 13.68715318 104.015249

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.