Abstract
The tribe Myonycterini comprises five fruit bat species of the family Pteropodidae, which are endemic to tropical Africa. Previous studies have produced conflicting results about their interspecific relationships.Here, we performed a comparative phylogeographic analysis based on 148 complete cytochrome b gene sequences from the three species distributed in West Africa and Central Africa (Myonycteris torquata, Lissonycteris angolensis and Megaloglossus woermanni). In addition, we investigated phylogenetic relationships within the tribe Myonycterini, using a matrix including 29 terminal taxa and 7235 nucleotide characters, corresponding to an alignment of two mitochondrial genes and seven nuclear introns. Our phylogenetic analyses confirmed that the genus Megaloglossus belongs to the tribe Myonycterini. Further, the genus Rousettus is paraphyletic, with R. lanosus, sometimes placed in the genus Stenonycteris, being the sister-group of the tribes Myonycterini and Epomophorini. Our phylogeographic results showed that populations of Myonycteris torquata and Megaloglossus woermanni from the Upper Guinea Forest are highly divergent from those of the Congo Basin Forest. Based on our molecular data, we recommended several taxonomic changes. First, Stenonycteris should be recognized as a separate genus from Rousettus and composed of S. lanosus. This genus should be elevated to a new tribe, Stenonycterini, within the subfamily Epomophorinae. This result shows that the evolution of lingual echolocation was more complicated than previously accepted. Second, the genus Lissonycteris is synonymised with Myonycteris. Third, the populations from West Africa formerly included in Myonycteris torquata and Megaloglossus woermanni are now placed in two distinct species, respectively, Myonycteris leptodon and Megaloglossus azagnyi sp. nov. Our molecular dating estimates show that the three phases of taxonomic diversification detected within the tribe Myonycterini can be related to three distinct decreases in tree cover vegetation, at 6.5–6, 2.7–2.5, and 1.8–1.6 Ma. Our results suggest that the high nucleotide distance between Ebolavirus Côte d’Ivoire and Ebolavirus Zaire can be correlated with the Plio/Pleistocene divergence between their putative reservoir host species, i.e., Myonycteris leptodon and Myonycteris torquata, respectively.