Structure of African elephant populations

The structure of elephant populations from east and south Africa has been analyzed by Georgiadis et al. (1994) on the basis of restriction site variation of mitochondrial DNA. They used F statistics based on identity by descent in tests for subdivision and reached the conclusion that there was a significant differentiation at the continental level, but that "populations were not significantly subdivided at the regional levels." The data were reanalyzed by Monte-Carlo permutation tests where population subdivision was tested by using F statistics based on partitioning the total haplotype diversity among populations. This resulted in identical conclusions at the continental level, but revealed in addition a significant subdivision at the regional level indicating haplotype frequency differences among the populations.