On the hybridisation between two distantly related Asian turtles (Sacalia × Mauremys) SALAMANDRA 41 1/2 21-26 Rheinbach, 20 May 2005 ISSN 0036-3375 On the hybridisation between two distantly related Asian turtles (Testudines: Sacalia × Mauremys) JAMES R. BUSKIRK, JAMES F. PARHAM & CHRIS R. FELDMAN Abstract. This is the first report of a hybridisation between Sacalia and Mauremys (Bataguridae). New data and a review of the literature show that 19 batagurid hybridisations are documented. Many more undoubtedly exist, but have not been documented and reported. Most hybrids are members of the Cuora + Mauremys clade and 17 of 19 reported hybrids have at least one member from this clade. The Sacalia × Mauremys hybridisation reported here is only the third hybridisation between a species of the Cuora + Mauremys clade and a species outside of that clade. Key words. Bataguridae: Mauremys, Sacalia; hybrids. Introduction that two possible solutions (lumping or split- ting) can provide a monophyletic taxonomy. Hybrids of batagurid (Bataguridae = Geo- We support and use the “lumping” solution emydidae; see JOYCE et al. 2004) turtles were following the arguments of FELDMAN & PAR- first reported from Japan where researchers HAM (2002, 2004), PARHAM & FELDMAN (2002), discovered the propensity for distantly re- and SPINKS et al. (2004). The hybrids reported lated species to breed in the wild and in here are between two distantly related bata- captivity (AOKI 1990, YASUKAWA et al. 1992, gurid species, Sacalia quadriocellata and OTANI 1995a, b). Since that time, several other Mauremys reevesii. We take this opportunity batagurid hybrids have been identified (FRITZ to review all documented batagurid hybrids. 1995, WINK et al. 2001, PARHAM & SHI 2001, PARHAM et al. 2001, SHI & PARHAM 2001, FRITZ & MENDAU 2002, GALGON & FRITZ 2002, SCHIL- Materials and Methods DE et al. 2004, SPINKS et al. 2004), some of which match species described from the The hybrid turtles reported here were Hong Kong pet trade (WINK et al. 2001, PAR- hatched by a private breeder in California, HAM et al. 2001, SPINKS et al. 2004). New USA, in August of 2000. Of approximately hypotheses of batagurid evolutionary rela- 300 terrestrial, semi-aquatic, and aquatic tur- tionships (BARTH et al. 2004, SPINKS et al. tles kept by this breeder, the majority live 2004) allow us to compare these hybridisa- indoors in large aquaria, cattle troughs, or tion events in a phylogenetic context. Most custom-made enclosures. Among the latter hybrids occur within or between the genera are several water-tight plexiglass enclosures Cuora and Mauremys. We use the latter ge- measuring 60 × 60 × 240 cm. The hybrids nus to include Chinemys and Ocadia be- were created in one of these custom enclo- cause multiple groups of workers have dis- sures filled with water to a depth of approxi- covered that Chinemys and Ocadia are nes- mately 15 cm. In the spring of 2000, this ted within traditional Mauremys (FELDMAN & enclosure included the following turtles (all PARHAM 2004, BARTH et al. 2004, SPINKS et al. of unknown origin, purchased via the pet 2004). BARTH et al. (2004) correctly point out trade): six Mauremys reevesii (four females), © 2005 Deutsche Gesellschaft für Herpetologie und Terrarienkunde e.V. (DGHT) http://www.salamandra-journal.com 21 JAMES R. BUSKIRK et al. five Sacalia quadriocellata (three females), fragment of the mitochondrial gene COI and and four Mauremys caspica caspica (two a 900 bp fragment of the mtDNA gene ND4 females). Prior to 2000, only the M. reevesii and linked tRNAs from specimen MVZ had successfully reproduced under the bree- 241502 and compared it to the larger data set der’s care. The breeder assumed genetic in- that includes every valid species of Maure- compatibility among the species and no ag- mys (FELDMAN & PARHAM 2004; see also for gressive behaviour was noted. However, the sequencing methods). breeder noticed that the male S. quadriocel- lata exhibited rather indiscriminate sexual activity. Results For nesting purposes, as well as for aerial Molecular data basking, a rubberised plastic tub measuring about 60 × 45 × 12 cm was kept filled with The sequence data of MVZ 241502 (Gen- sand and mulch, reached by a ramp from the bank AY562183; AY562185) are almost i- water below. Concrete blocks at the opposite dentical (< 0.6 % sequence divergence) to end provided additional basking areas. The those from a pet trade specimen of Mauremys breeder did not witness oviposition by any reevesii (MVZ 230533; Genbank AF turtle during the spring of 2000 but discov- 348263, AF348288), but are distinct from all ered two clutches, consisting of four and five other Mauremys (5.6 % different from M. elongated eggs, in June. These eggs resem- caspica). Thus, the DNA data confirm that bled those of M. reevesii yet were more lat- one (or more) of the female Mauremys reeve- erally compressed and slightly less oval. sii in the enclosure, not the Mauremys caspi- The eggs hatched after incubating in ca, laid the nine eggs. The presence of “eye moist vermiculite (1.5 parts water to 1 part spots” on the back of the head (Fig. 1) and vermiculite) at a temperature of 28 °C after other Sacalia features (see below) implicate 70-75 days. The hatchlings did not resemble that the father was one of the aforementioned any neonate turtles previously seen by the Sacalia quadriocellata that exhibited indis- breeder nor JRB. Because the neonates had criminant sexual activity. head spots as in Sacalia, but did not match Sacalia in other features, they were recog- nised as hybrids with either M. caspica or M. Description of hybrids reevesii. The nine hybrids have since been dis- The light brown carapace of the hybrids is persed to several collections. Three died of heavily marked with tiny black dots and the natural causes and were preserved in the periphery is highlighted by a pale streak on Museum of Vertebrate Zoology (MVZ), Ber- the outermost edges of most or all marginal keley, California, USA. Their specimen num- scutes. The plastron has a symmetrical or bers are MVZ 241500, 241501, and 241502. subsymmetrical, cloudy, dark, central figure, In Autumn 2003, one of the remaining six the peripheral third of the plastron scutes are specimens escaped from a private collection whitish (Fig. 1). The overall shape of the and is apparently lost forever. At the time of neonates is oval, not subcircular as seen in this writing, the remaining five are still alive many hatchling turtles. The soft parts are in private collections in the following Uni- predominantly grey, especially the head. ted States cities: Conyers, Georgia (JODY KAR- The flexor surfaces of the limbs are pale, and LIN, n=2); Oakland, California (JRB, n=1); the chin is heavily streaked. Some of the Berkeley, California (JFP, n=1); Davis, Cali- forelimb scales as well as tiny tubercles on fornia (PHIL S PINKS, n=1). the thighs are partly or completely whitish. In order to determine the maternal parent- Behind the eye is a prominent yellow age of the clutches, we sequenced a 700 bp spot, or sometimes paired spots, the smaller 22 On the hybridisation between two distantly related Asian turtles (Sacalia × Mauremys) Fig. 1. Sacalia quadriocellata × Mauremys reevesii hybrids. A) Living specimen in the care of JFP showing head coloration including detail of “eye spots” and malar region; B) same specimen in ventral view; C) The specimen sequenced in this study (MVZ 241502) in ventral view. lying between the eye itself and the larger Review of batagurid hybridisations temporal spot (Fig. 1). These “eye spots” are a diagnostic feature of Sacalia. On the top of Nineteen hybridisation events are documen- the head a yellow stripe extends from the ted for batagurid turtles, most from captivity. occiput toward the nuchal region, flanked by The authors have seen others in private col- a similar stripe on each of the temporal neck lections (see Discussion) or in Chinese turtle folds which did not quite reach the larger farms. However, in some cases the parental temporal spot itself. Beneath these, a poorly species can only be inferred. Our list is re- defined reticulate assemblage of yellowish stricted to previously reported hybrids from squiggles, some edged in black, adorn the the scientific literature and our study. The malar region. hybrids reported by AOKI (1990), FRITZ The “eye spots” on the back of the head (1995), FRITZ & MENDAU (2002), FRITZ & WI- clearly indicate that one of the parents must SCHUF (1997), GALGON & FRITZ (2002), OTANI have been a Sacalia. Although the tricari- (1995a, b), and STEMMLER (1973) are not ve- nate carapace, long tail, and pale iris are rified by genetic data. Where known, the reminiscent of neonate M. reevesii (Fig. 1), paternal species is listed first, maternal spe- the overall shape of the neonates is less cies second. In some cases, we do not know robust and more flattened. In this respect, the the specific identity of the maternal versus hybrids match the Sacalia neonates (Fig. 1). paternal lineages. We mark these hybridisa- The predominantly greyish soft parts of the tions with an asterisk (*). hybrids, particularly the head, are highly distinct from those of a neonate M. reevesii. Similarly, the chin of the hybrids is heavily a) Within Cuora (n=2) rather than sparsely streaked as in M. reeve- Cuora mouhotii × Cuora bourreti (“Cuora sii. Furthermore, the highly variable malar serrata” – PARHAM et al. 2001, STUART & pattern in neonate M. reevesii is never as PARHAM 2004). Origin: unknown, but probab- prominent or “garish” as in the hybrids. ly wild. 23 JAMES R. BUSKIRK et al. Cuora mouhotii × Cuora galbinifrons (“Cu- e) Both parental species outside of Cuora ora serrata” – PARHAM et al. 2001, STUART & + Mauremys clade (n=2) PARHAM 2004). Origin: unknown, but probab- Rhinoclemmys punctularia × Rhinoclemmys ly wild (see SHI et al.