The possibility that mosquitoes are diverted from protected people sleeping outdoors in a bednet to an unprotected person immediately adjacent was investigated during field trials conducted in an Afghan refugee village. The presence of an untreated bednet covering three people sleeping adjacent to the subject did not cause a significant rise in the biting rate on that subject. At the dose rates studied, biting rates in the presence of an ImperatorR treated bednet fell slightly (although not significantly). K-OthrineR treated bednets had no discernible effect on biting rates. Similar studies conducted within giant net traps showed that the ImperatorR treated bednet, and to a lesser extent the K-OthrineR treated bednet, did repel mosquitoes. Within these net traps, biting on the unprotected subject decreased considerably in the presence of the imperatorR treated bednets (P < 0.05 for both culicines and Anopheles stephensi Liston in the presence of the ImperatorR treated bednet). The net traps held in volatile components from the treated bednets in much the same way as the walls of a well ventilated room. We conclude that while sleeping next to a treated bednet may be of considerable benefit inside, where one sleeps in relation to a bednet outside is of little importance. Further studies revealed that for each mosquito species, insecticide induced mortality was highest when the bednet contained the preferred host of that species. This suggests that selection for pyrethroid resistance by bednets may be relatively slow to develop in areas where the primary vectors of malaria are zoophilic.