Deep genealogical lineages in the widely distributed African helmeted terrapin: evidence from mitochondrial and nuclear DNA (Testudines: Pelomedusidae: Pelomedusa subrufa).

We investigated the phylogeographic differentiation of the widely distributed African helmeted terrapin Pelomedusa subrufa based on 1503 base pairs of mitochondrial DNA (partial cyt b and ND4 genes with adjacent tRNAs) and 1937 bp of nuclear DNA (partial Rag1, Rag2, R35 genes). Congruent among different analyses, nine strongly divergent mitochondrial clades were found, representing three major geographical groupings: (1) A northern group which includes clades I from Cameroon, II from Ghana and Ivory Coast, III from Benin, Burkina Faso and Niger, IV from the Central African Republic, and V from Kenya, (2) a northeastern group consisting of clades VI from Somalia, and VII from Saudi Arabia and Yemen, and (3) a southern group comprising clade VIII from Botswana, the Democratic Republic of Congo, Madagascar and Malawi, and clade IX from South Africa. Malagasy and continental African populations were not clearly differentiated, indicating very recent arrival or introduction of Pelomedusa in Madagascar. The southern group was in some phylogenetic analyses sister to Pelusios, rendering Pelomedusa paraphyletic with respect to that genus. However, using partitioned Bayesian analyses and sequence data of the three nuclear genes, Pelomedusa was monophyletic, suggesting that its mitochondrial paraphyly is due to either ancient introgressive hybridization or phylogenetic noise. Otherwise, nuclear sequence data recovered a lower level of divergence, but corroborated the general differentiation pattern of Pelomedusa as revealed by mtDNA. This, and the depth of the divergences between clades, indicates ancient differentiation. The divergences observed fall within, and in part exceed considerably, the differentiation typically occurring among chelonian species. To test whether Pelomedusa is best considered a single species composed of deep genealogical lineages, or a complex of up to nine distinct species, we suggest a future taxonomic revision that should (1) extend the geographical sampling of molecular data, specifically focusing on contact zones and the possible sympatric occurrence of lineages without admixture, and (2) evaluate the morphology of the various genealogical lineages using the type specimens or topotypical material of the numerous junior synonyms of P. subrufa.