A survey was undertaken in the Darfur Region of Sudan, to collect retrospective mortality data for the famine period of 1984–5. During the famine the crude death rate trebled to 40 per thousand. Most of the excess mortality consisted of a sharp seasonal rise in child deaths. Infant and adult mortality rose only to a lesser extent. Fertility was lowered. The main causes of death were diarrhoeas, measles and malaria. Health factors, such as water supply, were associated with mortality differentials, and socio-economic variables such as wealth were not. It is argued that the excess mortality can be explained by severe localised health crises which caused increased exposure to life-threatening diseases. This conflicts with the established ‘starvation model’ of famine mortality, in which it is suggested that excess mortality occurs mainly through increased risk of death due to malnutrition.