Although it is well recognised that both behavioural and environmental factors play a role in determining small-scale heterogeneities in schistosomiasis transmission, empirical evidence of their relative importance is often limited. A study was conducted around Chaani, a village in northern Unguja (Zanzibar) where urinary schistosomiasis is endemic, in order to shed light upon the micro-epidemiology of Schistosoma haematobium and patterns of infection within schoolchildren and the intermediate host snail Bulinus globosus, that may help in adjusting recently implemented control programmes. Malacological surveys were conducted to identify transmission foci and questionnaires were devised to assess recalled water-contact patterns of 150 schoolchildren who had been screened for S. haematobium infection, finding an overall prevalence of 50.6%. Boys were more frequently and more heavily infected than girls and, accordingly, mean exposure scores were significantly higher for boys than girls. Univariate statistics revealed significant associations between infection and specific water-contact activities, including washing/bathing (odds ratio [OR]=3.01, 95% confidence interval [CI]=1.36-6.67) and playing (OR=4.03, 95% CI=1.51-10.70) in streams/ponds. In multivariate analysis, however, the strongest predictor of infection was proximity of the child's home to a site harbouring S. haematobium-infected B. globosus (P<0.001), suggesting that geography may be a better proxy for exposure than self-reported water contact. Surveillance programmes may therefore wish to prioritise evaluation of environmental risk to pinpoint transmission at the micro-geographical level, although water-contact questionnaires are also recommended as a complementary tool to rapidly identify the behaviour patterns of children at most risk of infection in the rural communities of Zanzibar. Such knowledge is a prerequisite for focusing and improving schistosomiasis control at the local level.