This study had 2 objectives: first, to investigate how the processes of slender form replication, of differentiation from dividing slender to non-dividing stumpy forms, and of stumpy mortality, combine to determine the initial (acute-phase) growth rate of Trypanosoma brucei populations; second, to determine how acute-phase growth rates influence parasite densities during the subsequent, chronic phase of infection. During the acute phase, slender and stumpy populations both grew approximately exponentially, the latter more slowly than the former. Mathematical models showed how this difference in slender and stumpy growth rates can be explained in terms of heterogeneous replication and differentiation rates. Stumpy life-expectancy was determined for one stock and found to be age-dependent with a half-life of 48–72 h, much larger than observed population doubling times of 5–10 h. A comparison of cloned stocks showed that the highest parasite densities during the chronic phase were associated with the highest acute-phase growth rates of both the whole parasite population and of the subpopulation of slender forms. By contrast, high chronic-phase parasitaemias artificially produced following rapid syringe passage were associated with low acute-phase growth rates of slender forms. Syringe-passaging is a laboratory procedure which selects for virulent parasites, but these parasites behave differently from naturally virulent stocks.