We compared the anti-mosquito and antimalarial potentialities of placebo-treated versus permethrin-impregnated bed nets in north-western Guinea Bissau. Baseline, pre-intervention entomological and parasitological data were collected during the rainy season of 1990 and bed nets were distributed shortly before the rainy season of 1991. Pairs of 3 ethnically different villages were investigated. The villages in each pair were at least 2 km apart but belonged to the same ethnic group in an ecologically similar area. After one year permethrin-treated bed nets were provided to all people in one village of each pair and placebo-treated bed nets to the other villages. About 98% of mosquitoes caught in bedrooms belonged to Anopheles gambiae and A. melas, which we consider to be the main malaria vectors in the study villages. Mean Plasmodium falciparum sporozoite rate in A. gambiae (9.6%) and A. melas (12.4%) was highest during October-November. The Plasmodium index in children 2-9 years old in the 6 villages, at the end of the rainy season 1990, ranged between 44% and 79%. Of these, 98% were identified as P. falciparum, 1% as P. malariae and 1% as mixed infections of these species. Significant reductions of Anopheles indoor resting densities and malaria parasite rates in humans were recorded in villages which had received permethrin-treated nets, but not in the control villages. The mean number of P. falciparum-infective mosquito bites received indoors in untreated villages during the rainy season was estimated to be about 4 per child and 20 per adult. This inoculation rate was reduced by at least 78% by the use of permethrin-impregnated bed nets. The malaria parasite rates and proportions of people experiencing 'disease with fever' decreased significantly in villages provided with permethrin-treated nets but not in the control villages. Impregnated nets may be an important tool to reduce disease and death due to malaria in Guinea Bissau.