African elephants are conventionally classified as a single species: Loxodonta africana (Blumenbach 1797). However, the discovery in 1900 of a smaller form of the African elephant, spread throughout the equatorial belt of this land, has given rise to a debate over the relevance of a second species of elephant in Africa. The twentieth century has not provided any definite answer to this question. Actually, recent molecular analyses have sustained this issue by advocating either a division of forest elephants into a valid species, or their inclusion as a subspecies of L. africana. Our work initiated at the National Museum of Natural History of Paris provides new molecular (mitochondrial) and morphological (and morphometrical) evidence making it possible to propose a comprehensive phylogenetic hypothesis. It appears that there is no conclusive argument to keep forest elephants (cyclotis form) and savannah elephants (africana form) apart in two distinct species. A high level of mitochondrial introgression between the two forms, as well as a continuum in the morphology of the skulls of the two morphotypes rather suggests that, despite an ancient division, these two taxa freely interbreed wherever their ranges intersect. We thus adopt a conservative systematic position in considering these two forms as two subspecies, respectively: L. africana africana, the savannah elephant, and L. africana cyclotis, the forest elephant. We finally discuss the conservation topic in the light of this systematic framework.