In forest and forest fringe settings of northwest Thailand, Anopheles minimus A was found to be the primary vector while An. dirus s./., An. sawadwongporni and An. maculatus s.s. are secondary vectors of malaria. The sites of transmission were investigated entomologically; parallel epidemiological investigations were made by another team. Malaria transmission in the villages in terms of the number of cases was as or more important than in the farm huts or in the forest, although the daily risk of infection was greater in forest activities. In a cage tunnel, lambdacyhalothrin on netting significantly reduced feeding success and killed mosquitoes which attempted to bite on a human arm through the treated nets. The aerial toxicity of the insecticide is apparently due to the spread of insecticide into the air as dust rather than as vapour. The biting rate of mosquitoes on humans close to a treated net was significantly reduced. The entomological impact of treated nets was evaluated in both long-term (two years) and short-term (48 days) studies in five and four communities, respectively. The results in the long-term evaluation showed that treated nets had very little effect on the densities and parous rates of An. minimus A, An. sawadwongporni and An. maculatus s.s. populations. However, the evaluation was carried out with a high washing rate, low coverage of re-impregnation and variations in climatic factors. In the short-term evaluation with intensive mosquito sampling, no mass effect on An. minimus A was observed. The exophagy, zoophily and early evening biting behaviours are probably the main factors reducing the effectiveness of treated nets. It is concluded that treated nets may provide improved personal protection especially among mobile populations but have little benefit in reducing mosquito population vectorial capacity of these species. There was cross-reactivity between the sporozoite monoclonal antibodies employed in the enzyme-linked immunosorbent assay and unknown factor(s) in blood of cows, buffaloes and pigs. This presumably caused false positive results when testing blood-fed zoophilic mosquitoes. Membrane feeding experiments showed that several zoophilic species are susceptible to both Plasmodium falciparum and P. vivax, except for a few species that are refractory to the former.