1 Museum of Zoology (Museum fu ¨r Tierkunde), Senckenberg Dresden, Germany; 2 LÕAssociation du Refuge des Tortues, Bessie`res, France Molecular phylogeny of Central and South American slider turtles: implications for biogeography and systematics (Testudines: Emydidae: Trachemys) Uwe Fritz 1 ,Heiko Stuckas 1 ,Mario Vargas-Ramı ´rez 1 ,Anna K. Hundsdo ¨ rfer 1 ,Je ´ro ˆ me Maran 2 and Martin Pa ¨ ckert 1 Abstract We analyse phylogeny, systematics and biogeography of slider turtles (Trachemys spp.) using sequence data of four mitochondrial genes (3242 bp) and five nuclear loci (3396 bp) of most South American and southern Central American taxa and representatives of northern Central American, West Indian and North American slider species (16 species and subspecies) and allied North American species (genera Chrysemys, Deirochelys, Graptemys, Malaclemys, Pseudemys). By applying maximum likelihood, relaxed molecular clock and ancestral range analyses, we provide evidence for two successive colonizations of South America by slider turtles. In addition, we show that the current species delineation of Central and South American slider turtles is incorrect. Our data suggest that Trachemys grayi is a distinct polytypic species that embraces, besides the nominotypical subspecies, T. g. emolli and T. g. panamensis. Trachemys ornata is also polytypic with the subspecies T. o. ornata, T. o. callirostris, T. o. cataspila, T. o. chichiriviche and T. o. venusta. Moreover, T. adiutrix should be regarded as a subspecies of T. dorbigni. All studied Trachemys species are inferred to have originated in the Late Miocene to Early Pliocene. The ancestor of the two subspecies of T. dorbigni colonized South America most probably prior to the establishment of the land bridge connecting Central and South America, whereas the two South American subspecies of T. ornata represent a younger independent immigration wave from Central America. Key words: Ancestral range analysis – phylogeography – molecular clock – numt – species delineation – Chrysemys – Deirochelys – Graptemys – Malaclemys – Pseudemys Introduction The speciose slider turtle genus Trachemys Agassiz, 1857 represents one of the most widely distributed American reptile groups. Several parapatric species range over large parts of North and Central America to northern South America (Colombia and Venezuela); two further, widely disjunct South American species occur in northern Brazil (state of Maranha˜ o) and in the region of the Rio de la Plata (Brazil, Argentina, Uruguay). Four additional species live in the West Indies (Seidel 1988, 2002; Fritz and Havasˇ 2007; Rhodin et al. 2010; Fig. 1; Table 1). All Trachemys species are highly aquatic, medium- to large-sized turtles that leave the water only for basking, short overland movements and nesting (Gibbons 1990; Gibbons et al. 1990; Ernst et al. 2000; Ernst and Lovich 2009). The best-known taxon is the North American red-eared slider, Trachemys scripta elegans, which was traded for decades in enormous numbers as a pet. As a consequence, it has been naturalized worldwide in many regions outside its native range (Ernst et al. 2000; Fritz and Havasˇ 2007). Trachemys belongs to the mainly Nearctic family Emydidae and is its only genus that colonized Central and South America and the West Indies. Since the land connection of the two Americas did not form before the Pliocene (approximately 3 million years ago; Marshall 1979, 1985, 1988; Lessios 2008; Molnar 2008; Woodburne 2010), Trachemys has been postulated to be a recent invader of Central and South America (Savage 1966; Pritchard and Trebbau 1984; Vanzolini and Heyer 1985; Legler 1990; Moll and Moll 1990; Seidel and Jackson 1990; Vanzolini 1995; de la Fuente et al. 2002). Slider turtles generally do not occur in closed forest or rainforest, probably because of the difficulty in finding suitable, open nesting sites (Moll and Legler 1971; Pritchard and Trebbau 1984; Moll and Moll 1990). Accordingly, the South American species are confined to regions north and south of the Amazon Basin. This conspicuous distribution pattern suggests that the two disjunct southernmost species, T. adiutrix and T. dorbigni, might have crossed the Amazon Basin during a phase of rainforest fragmentation (Pritchard and Trebbau 1984; Moll and Moll 1990) and that the northern South American T. callirostris is their closest relative (Wil- liams 1956). If this scenario should be correct, it would strongly support the contentious Ôforest refugia hypothesisÕ. According to it, increasing aridity and cooling during the Pleistocene glaciations resulted in isolated rainforest patches being separated by savannahs and deserts (e.g. Haffer 1969, 1997; Prance 1973; Potts and Behrensmeyer 1992; Hooghiem- stra and van der Hammen 1998). This hypothesis was widely accepted until a few years ago, but has been severely criticized recently (e.g. Kastner and Gon˜i 2003; Pennington et al. 2004) because pollen cores show no evidence of reduced rainforest cover in the Amazon Basin (e.g. Colinvaux et al. 2001; Mayle et al. 2004), dynamic vegetation model simulations reject the hypothesis of widespread grasslands in Amazonia during the last glacial maximum (e.g. Cowling et al. 2001) and genetic data suggest that diversification in tropical rainforest animals generally predates the Pleistocene (Moritz et al. 2000; Glor et al. 2001). Yet, the genetic structure of two open-habitat reptiles argues in favour of former forest fragmentation (Chelonoidis carbonaria: Vargas-Ramı´rez et al. 2010a; Crotalus durissus: Wu¨ster et al. 2005; Quijada-Mascaren˜as et al. 2007). At least in the case of Ch. carbonaria, the onset of the vicariant event corresponding to forest fragmentation seems to predate Corresponding author: Uwe Fritz (uwe.fritz@senckenberg.de) Contributing authors: Heiko Stuckas (heiko.stuckas@senckenberg.de), Mario Vargas-Ramı´rez (mario.vargas@senckenberg.de), Anna K. Hundsdo¨rfer (anna.hundsdoerfer@senckenberg.de), Je´roˆme Maran (jerome.maran@wanadoo.fr), Martin Pa¨ckert (martin.paeckert@ senckenberg.de) Ó 2011 Blackwell Verlag GmbH Accepted on 13 October 2011 J Zool Syst Evol Res doi: 10.1111/j.1439-0469.2011.00647.x J Zool Syst Evol Res (2012) 50(2), 125–136