The epidemiology of rabies in dogs in Zimbabwe is described using data from 1950, when rabies was re-introduced after a 37-year absence, to 1996. Dogs constituted 45.7% of all laboratory-confirmed rabies cases and were the species most frequently diagnosed with the disease. Slightly more cases were diagnosed from June to November than in other months. From 1950 to the early 1980s, most dog cases were recorded from commercial farming areas, but since the early 1980s most have been recorded from communal (subsistence farming) areas. This change appears to be due to improved surveillance in communal areas and not to any change in the prevalence of rabies. Dog rabies therefore appears to be maintained mainly in communal area dog populations, particularly the large communal area blocks. Urban rabies was not important except in the city of Mutare. Where dog rabies prevalence was high, the disease was cyclic with periods between peak prevalence ranging from 4-7 years. Dog rabies cases were, on the whole, independent of jackal rabies and rabies in other carnivores. There was a significant negative relationship between the annual number of rabies vaccine doses administered nationally to dogs and the annual number of dog rabies cases lagged by one year, indicating that the past levels of immunisation coverage have had a significant effect on the number of rabies cases. However, dog vaccination coverage has clearly not been adequate to prevent the regular occurrence of rabies in dogs.