OBJECTIVES--To identify which British residents travelling abroad are at greatest risk of malaria infection, and to determine the efficacy of malaria chemoprophylaxis for preventing P falciparum infections in tropical Africa. DESIGN--Prospective cohort study (case-base linkage) with routine national surveillance systems. Denominators (base population) were obtained from monitoring a random sample of returning British travellers with the international passenger survey. Numerators (cases) were obtained from reports of malaria infections in British residents, through the Malaria Reference Laboratory network. SETTING--International passenger survey conducted at passport control of international airports in Britain. Malaria reports received nationally were collated centrally in London. SUBJECTS--2948 British residents (0.2%) returning to Britain in 1987 randomly selected and questioned and 1052 British residents with microscopically confirmed malaria infections in 1987, whose case reports were reviewed and on whom additional data were collected by postal survey. MAIN OUTCOME MEASURES--Annual incidence subdivided by categories of risk. Chemoprophylactic efficacy for east and west Africa by principal regimens and compliance. RESULTS--Annual rates of reported infection per 100,000 travellers to Oceania were 4100; to west and east Africa were 375 and 172 respectively; to Latin America, the Far East, and the Middle East were 12, 2, and 1 respectively. Immigrants visiting friends and relatives in Ghana and Nigeria were at greatest risk (1303 and 952 per 100,000 respectively) in west Africa. Business travellers to Kenya experienced the highest attack rates in east Africa (465 per 100,000). Age-sex specific attack rates varied by region. No prophylaxis was reported to have been used by 23% of British visitors to west Africa, 17% to east Africa, 46% to central or southern Africa, and 58% visiting south Asia. The efficacy of chloroquine plus proguanil against P falciparum infection was 73% and 54% in west and east Africa respectively. Lower values were obtained for chloroquine alone and proguanil alone. The efficacy of Maloprim (pyrimethamine-dapsone) was 61% in west Africa, but only 9% in east Africa. Visitors to west Africa who did not comply with their chemoprophylactic regimen were at a 2.5-fold higher risk of infection than fully compliant users. Non-compliant visitors to east Africa had similar rates of infection as non-drug users. CONCLUSIONS--In 1987 chloroquine plus proguanil was the preferred chemoprophylactic regimen for P falciparum infection in Africa; antimalarial drugs must be taken regularly to be effective.