A field trial of permethrin-impregnated bed nets (PIBs) was conducted in 2 Afghan refugee villages in Pakistan. Nets were issued to only 10% of families (= 1398 people); this simulated a situation in which bed nets are gradually adopted by villagers in Afghanistan. A further 10% lacking bed nets were selected as controls from the same villages. An initial survey showed that 86% of household heads were aware that malaria was transmitted by mosquito bites, but only 2% had used bed nets before. Trial families were encouraged to attend the village health centres if they fell ill. Microscopy records showed that, between July and December 1991, 22.4% of the control group became infected with Plasmodium vivax and 13.0% contracted P. falciparum while in the intervention group only 9.9% contracted P. vivax (relative risk 0.58, confidence interval [CI] 95% 0.49-0.68) and only 3.8% contracted P. falciparum (relative risk 0.39, 95% CI 0.29-0.53). A single treatment of the nets with permethrin at 0.5g/m2 remained protective throughout the 6 months' transmission season. 73% of families claimed to use their nets every night; members of families who claimed to use nets less regularly showed an incidence similar to that of the control group. There was no sex or age difference in net use or protective efficacy. Headlouse infestation rates were reduced in PIB users. Few nets were washed, given away or sold. The prospect for PIBs as personal protection appears good, despite people's lack of previous experience.