Mitochondrial DNA control region sequence variation was obtained and the population history of the common hippopotamus was inferred from 109 individuals from 13 localities covering six populations in sub-Saharan Africa. In all, 100 haplotypes were defined, of which 98 were locality specific. A relatively low overall nucleotide diversity was observed (p ¼ 1.9%), as compared to other large mammals so far studied from the same region. Within populations, nucleotide diversity varied from 1.52% in Zambia to 1.92% in Queen Elizabeth and Masai Mara. Overall, low but significant genetic differentiation was observed in the total data set (F ST ¼ 0.138; P ¼ 0.001), and at the population level, patterns of differentiation support previously suggested hippopotamus subspecies designations (F CT ¼ 0.103; P ¼ 0.015). Evidence that the common hippopotamus recently expanded were revealed by: (i) lack of clear geographical structure among haplotypes, (ii) mismatch distributions of pairwise differences (r ¼ 0.0053; P ¼ 0.012) and site-frequency spectra, (iii) Fu's neutrality statistics (F S ¼ À155.409; Po0.00001) and (iv) Fu and Li's statistical tests (D * ¼ À3.191; Po0.01, F * ¼ À2.668; P ¼ 0.01). Mismatch distributions, site-frequency spectra and neutrality statistics performed at subspecies level also supported expansion of Hippopotamus amphibius across Africa. We interpret observed common hippopotamus population history in terms of Pleistocene drainage overflow and suggest recognising the three subspecies that were sampled in this study as separate management units in future conservation planning. Heredity (2005) 95, 206-215.