Pyrethroid-impregnated bednets are advocated for personal protection against malaria vectors. To avoid the need for periodic re-treatment, it would be advantageous to have nets that retain insecticidal efficacy for years and withstand repeated washing. Such a type of commercially produced bednet with permethrin 2% incorporated in polyethylene fibres (trademark Olyset Net supplied by Sumika Life-Tech Co., Osaka, Japan) was evaluated against mosquitoes in veranda-trap huts at Yaokoffikro, near Bouaké, C te d'Ivoire, by standard WHOPES phase II procedures. Four Olyset Nets were compared with a standard untreated polyester net as control. They comprised three examples previously used in a village for over 3 years (one washed, one dirty, one very dirty) and a previously unused Olyset Net, newly unwrapped, from the same original batch. Bioassays with 3 min exposure of susceptible Anopheles gambiae Giles (Diptera: Culicidae) gave >99% mortality of female mosquitoes tested on the 'new' Olyset Net. The used Olyset Nets gave mortality rates averaging 83% for the washed net, 85% for the dirty net and 55% for the very dirty net (within 24-h following 3 min exposure). Thus, Olyset Nets were found to remain remarkably effective against susceptible An. gambiae for at least 3 years under field conditions. Wild pyrethroid-resistant populations of Culex quinquefasciatus Say and An. gambiae (savanna cytotype with 96% kdr) were assessed during June-August 1999 for their responses to sleepers protected by nets in the experimental huts. With regard to hut entry by foraging female mosquitoes, Olyset Nets showed some deterrency against An. gambiae (44% reduction by the new net, approximately 20% by the dirty nets, none by the washed net), but not against Cx. quinquefasciatus. Among mosquitoes entering the hut with untreated control net, 30-34% tried to leave (exophily) but were caught in the verandah trap. The permethrin repellency of Olyset Nets increased exophily by 19% for An. gambiae and 14% for Cx. quinquefasciatus. Blood-feeding rates were 16% An. gambiae and 35% Cx. quinquefasciatus in the hut with sleeper under the untreated net (showing considerable prevention of biting), 22-26% of both species in huts with washed or dirty used Olyset Nets (not significantly different from control), while the biting success rate of Cx. quinquefasciatus (but not kdr An. gambiae) was more than halved by the 'new' Olyset Net. Mortality rates of pyrethroid-resistant An. gambiae and Cx. quinquefasciatus from the huts were, respectively, 3% and 8% with the untreated polyester net, 27.5% and 17% with the 'new' Olyset, 15% and 17.5% with the washed Olyset, 16-25% and 17-20% with dirty old Olyset Nets. Kill differences between nets are significantly different for both An. gambiae and Cx. quinquefasciatus. Unfortunately the washed used Olyset Net showed least activity against resistant mosquitoes, despite its greatest activity against susceptible An. gambiae. In each case there was evidence that a high proportion of mosquitoes failed to feed through the net (many of them dying from starvation when they could not leave the closed hut), with indications that dirty Olyset nets enhanced this protective value.