Joint sequences from the mitochondrial cytochrome b and 16S rRNA genes of a
wide representation of Megachiroptera were employed to evaluate the tradit
ional taxonomic arrangement of African fruitbats and to examine their origi
ns and evolutionary relationships. The resulting phylogenetic hypotheses ar
e inconsistent with the previously established morphology-based subdivision
s of Megachiroptera at the suprageneric level. Findings indicate the existe
nce of an African clade, which appears to be formed by two endemic clades:
the epomophorines and the myonycterines. According to our topologies, the g
enus Rousettus is monospecific in mainland Africa. Its traditional subgener
a Stenonycteris and Lissonycteris appear closer to the myonycterines than t
o Rousettus. Topologies also indicate that the African genus Eidolon is not
phylogenetically related to any other African fruitbat. It would seem that
the arrival of fruitbats in Africa was a complex process involving at leas
t three independent colonization events. One event took place probably in t
he Miocene via forested corridors that connected. the African and Asian rai
n forest blocks, as for other groups of mammals. The resulting lineage dive
rsified into most of the extant African fruitbats. Related to this clade, t
he Rousettus species group is thought to have arrived in Africa in more rec
ent times, possibly by progressive displacement from the East through India
. Finally, the present topologies suggest an independent colonization of Af
rica by ancestors of Eidolon. (C) 1999 Academic Press.