As part of a study of malaria morbidity an the effectiveness of treated nets in areas that were expected to differ greatly in intensity of malaria transmission, mosquitoes were collected from light traps set in both lowland and highland hamlets in the Muheza district, Tanzania. They were sent to London and by using the ELISA method the CSP/sporozoite positivity rates and hence the entomological inoculation rates (EIRs) were determined. The sporozoite rates were significantly lowered by the introduction of alphacypermethrin-treated nets. The EIRs (per year) for the control and treated hamlets were estimated to be 397 and 104 respectively in the lowlands and 16.6 and 5.7 respectively in the highlands. Thus the overall impact of the nets on EIR for someone outside a net was a reduction of 74% in the lowlands and 66% in the highlands. These reduction factors are not as great as in earlier trials in the same area. The EIR in the lowlands was much greater than in the highlands purely because of the larger numbers of mosquitoes in the former: the sporozoite rates did not differ significantly.