There is much confusion over which malaria chemoprophylaxis should be used in areas such as East Africa. We did two consecutive studies between 1985 and 1991 to assess the efficacy and side-effects of malaria chemoprophylaxis in short-term travellers to East Africa. All passengers returning from Kenya to Europe received an in-flight questionnaire and a second one three months later. Any report of documented malaria or of admission to hospital for possible side-effects was verified with the physician. 145 003 travellers completed questionnaires. Among the 139 164 who stayed in East Africa for less than one year, 296 cases of confirmed malaria were reported (275 due to P falciparum). In people who used no chemoprophylaxis, the incidence of falciparum malaria was 1.2% per month. Prophylactic effectiveness was 91% (95% Cl 85 to 94) for mefloquine, 82% (71 to 89) for pyrimethamine and sulfadoxine, 72% (56 to 82) for chloroquine plus proguanil, and 10 to 42% for chloroquine at various doses. Rates of side-effects, which were usually mild, were 18.8% for mefloquine users, 17.1% and 18.6% for chloroquine 300 mg and 600 mg base per week, respectively, 30.1% for chloroquine plus proguanil, and 11.7% for sulfadoxine and pyrimethamine. Mefloquine is significantly more effective than chloroquine plus proguanil for malaria prophylaxis in short-term tourists visiting East Africa and has a tolerance similar to that of chloroquine used alone.