Online Supplementary Material- Supplemental Text

Comparative population genetics of aquatic turtles in the desert

S. E. McGaugh

Morphological identification of turtles in the basin

Trachemys taylori was morphologically distinguished from congeners by two main criteria: 1) carapacial markings are ocellar and 2) plastral patterning is extensive, black, and expands from the seams along the entire width of plastron (Legler 1960). In the invasive Trachemys scripta, carapacial markings are largely longitudinal, and plastral patterns are well-defined spots and smudges in the center of each plastral scute. Foreclaws are not elongated in males of T. taylori, but foreclaws are elongated in T. scripta.

Three characteristics were used in identifying putative A. atra: 1) ‘‘blackish’’ pigmentation of dorsal surface; 2) speckled pigmentation on ventral surfaces; and 3) longitudinal corrugations on the posterior of the carapace (Webb and Legler 1960; Winokur 1968). In contrast, the invasive A. spinifera emoryi is defined by 1) tan to olive-brown carapace; 2) no ventral pigmentation; 3) defined pale marginal band; 4) white, raised tubercles on the back third of the carapace; and 5) a clear triangular facial pattern (Ernst and Lovich 2009) (Figure S1). Apalone atra also has more ovoid shaped adult carapaces and fainter marginal bands than A. s. emoryi and posterior white tubercles on the carapace of males only (Webb and Legler 1960; Winokur 1968); these characteristics were mainly used as secondary identifiers. Morphological identification of putative Apalone atra from A. s. emoryi has been described in further depth elsewhere (McGaugh and Janzen 2008; Winokur 1968).

Habitat preference between A. atra and A. spinifera emoryi in the basin putatively splits along species delimitations (Webb and Legler 1960; Winokur 1968); the darker, endemic A. atra prefers lagoons, and the light A. s. emoryi prefers rivers and playa lakes (Webb and Legler 1960; Winokur 1968). Lagoons have darker soil bottoms than rivers and playa lakes (McGaugh 2008), and pigmentation of the Apalone carapace appears to be strongly correlated with the coloration of the soil of particular aquatic habitats in Cuatro Ciénegas (McGaugh 2008).

Hypothesized history of turtles in the basin

Cytochrome b sequence data from (Spinks et al. 2009) was used to analyze how divergent T. taylori and T. coahuila are from their respective sister taxa to elucidate the timing of cladogenesis for these two endemics. For synonymous sites of cytochrome b, T. taylori is about 18.30% divergent from T. s. elegans (49/259 sites). Likewise, for synonymous sites of cytochrome b, T. coahuila is 19.75% divergent (47/271 sites) from its sister taxa, T. carolina. This concordance between synonymous substitution rate between each endemic and its sister taxa suggests that these two endemics speciated at similar time (if neutral molecular rate of evolution is assumed to be similar between the two species). Comparisons from a sample of two nuclear genes, reveal less concordance between estimates (RELN, intron only: Trachemys: 0.41% of 1008bp, T. coahuila: 0.10% of 1009bp; RAG-1: Trachemys 1.9% of 158 synonymous sites, Terrapene: 0 % of 169 synonymous sites; Spinks et al. 2009); thus, this conclusion must be taken as tentative.

In light of the similar divergence for T. taylori and T. coahuila and their respective sister taxa, the lack of genetic differentiation between Apalone atra and A. spinifera from the Rio Grande (McGaugh and Janzen 2008) suggests two hypotheses which may be potentially plausible and difficult to tease apart: 1) Apalone was not in the basin until recently, perhaps invading during the last wet period that connected the basin to the Río Salado de los Nadadores drainage, 2) Apalone has been in the basin as long as the other two species, but geneflow with Apalone outside of the basin occurred more often, perhaps, anytime an above ground water connection existed between the basin and the Río Salado de los Nadadores drainage.

1