This study presents new findings concerning the evolution of the human pathogens, Trypanosoma brucei and T. cruzi, which suggest that these parasites have divergent origins and fundamentally different patterns of evolution. Phylogenetic analysis of 18S rRNA sequences places T. brucei in a clade comprising exclusively mammalian trypanosomes of African origin, suggesting an evolutionary history confined to Africa. T. cruzi (from humans and sylvatic mammals) clusters with trypanosomes specific to Old and New World bats, T. rangeli and a trypanosome species isolated from an Australian kangaroo. The origins of parasites within this clade, other than some of those from bats, lie in South America and Australia suggesting an ancient southern super-continent origin for T. cruzi, possibly in marsupials; the only trypanosomes from this clade to have spread to the Old World are those infecting bats, doubtless by virtue of the mobility of their hosts. Viewed in the context of palaeogeographical evidence, the results date the divergence of T. brucei and T. cruzi to the mid-Cretaceous, around 100 million years before present, following the separation of Africa, South America and Euramerica. The inclusion in this study of a broad range of trypanosome species from various different hosts has allowed long phylogenetic branches to be resolved, overcoming the limitations of many previous studies. Moreover, T. brucei and the other mammalian tsetse-transmitted trypanosomes appear, from these data, to be evolving several times faster than T. cruzi and its relatives.