One African baobab species or two?: Synonymy of Adansonia kilima and A. digitata

We assessed the validity of a recently described baobab species Adansonia kilima that was suggested to be a diploid occurring in both eastern and southern Africa at high elevations within the range of the well-known tetraploid species A. digitata. We used a combination of phylogenetic analyses and statistical comparisons of various traits (e.g., flowers, stomata, pollen, chromosome counts) to test for the presence of two continental African baobab species. Ordination of the floral features of 133 herbarium specimens from across the natural range of A. digitata, including the putative type of A. kilima and other Tanzanian accessions as previously assigned A. kilima, revealed no distinct clusters of specimens. Likewise, stomatal size and density varied greatly across the specimens examined, with no clear bimodal pattern and no obvious association with altitude. The type specimen of A. kilima was found to have a chromosome number of 2n ≈ 166, showing it to be a tetraploid, like A. digitata. Phylogenetic analysis of the ITS region showed little resolution within the African baobab clade and a lack of distinction between the A. kilima type and A. digitata regional accessions. Among the 13 haplotypes detected, no distinct haplotype representing A. kilima was identified. Based on the data at hand we conclude that A. kilima is neither cytologically nor morphologically distinct and is here reduced to synonymy with A. digitata.