Knowledge on faunal diversification in African rainforests remains scarce. We used phylogeography to assess (i) the role of Pleistocene climatic oscillations in the diversification of the African common pangolin (Manis tricuspis) and (ii) the utility of our multilocus approach for taxonomic delineation and trade tracing of this heavily poached species. We sequenced 101 individuals for two mitochondrial DNA (mtDNA), two nuclear DNA and one Y-borne gene fragments (totalizing 2602 bp). We used a time-calibrated, Bayesian inference phylogenetic framework and conducted character-based, genetic and phylogenetic delineation of species hypotheses within African common pangolins. We identified six geographic lineages partitioned into western Africa, Ghana, the Dahomey Gap, western central Africa, Gabon and central Africa, all diverging during the Middle to Late Pleistocene. MtDNA (cytochrome b + control region) was the sole locus to provide diagnostic characters for each of the six lineages. Tree-based Bayesian delimitation methods using single- and multilocus approaches gave high support for 'species' level recognition of the six African common pangolin lineages. Although the diversification of African common pangolins occurred during Pleistocene cyclical glaciations, causative correlation with traditional rainforest refugia and riverine barriers in Africa was not straightforward. We conclude on the existence of six cryptic lineages within African common pangolins, which might be of major relevance for future conservation strategies. The high discriminative power of the mtDNA markers used in this study should allow an efficient molecular tracing of the regional origin of African common pangolin seizures.

The bushmeat trade in tropical Africa represents illegal, unsustainable off-takes of millions of tons of wild game - mostly mammals - per year. We sequenced four mitochondrial gene fragments (cyt b, COI, 12S, 16S) in >300 bushmeat items representing nine mammalian orders and 59 morphological species from five western and central African countries (Guinea, Ghana, Nigeria, Cameroon and Equatorial Guinea). Our objectives were to assess the efficiency of cross-species PCR amplification and to evaluate the usefulness of our multilocus approach for reliable bushmeat species identification. We provide a straightforward amplification protocol using a single 'universal' primer pair per gene that generally yielded >90% PCR success rates across orders and was robust to different types of meat preprocessing and DNA extraction protocols. For taxonomic identification, we set up a decision pipeline combining similarity- and tree-based approaches with an assessment of taxonomic expertise and coverage of the GENBANK database. Our multilocus approach permitted us to: (i) adjust for existing taxonomic gaps in GENBANK databases, (ii) assign to the species level 67% of the morphological species hypotheses and (iii) successfully identify samples with uncertain taxonomic attribution (preprocessed carcasses and cryptic lineages). High levels of genetic polymorphism across genes and taxa, together with the excellent resolution observed among species-level clusters (neighbour-joining trees and Klee diagrams) advocate the usefulness of our markers for bushmeat DNA typing. We formalize our DNA typing decision pipeline through an expert-curated query database - DNA BUSHMEAT - that shall permit the automated identification of African forest bushmeat items.