Abstract This paper reviews results of several ethnographic studies that have examined the issue of local terminology for malaria in Africa, then presents findings from an on-going study in Bagamoyo District, Tanzania. The study used a mixture of qualitative and quantitative interview methods to examine local perceptions of malaria and malaria treatment practices. Although the local term homa ya malaria or malaria fever appeared on the surface to correspond closely with the biomedical term malaria, significant and often subtle differences were found between the two terms. Of perhaps greatest importance, common consequences of malaria in endemic areas such as cerebral malaria in young children, severe anaemia and malaria in pregnancy were not connected with homa ya malaria by many people. A set of guidelines are described that were used to determine how best to promote acceptance and use of insecticide-impregnated mosquito nets, given these results. It is demonstrated that the position of the term used to denote malaria in the local taxonomy of febrile illnesses has important implications for the design of health education interventions.