The authors examined the relation between 50 widely used agricultural pesticides and lung cancer incidence in the Agricultural Health Study, a prospective cohort study of 57,284 pesticide applicators and 32,333 spouses of farmer applicators with no prior history of lung cancer. Self-administered questionnaires were completed at enrollment (1993-1997). Cancer incidence was determined through population-based cancer registries from enrollment through December 31, 2001. A lung cancer standardized incidence ratio of 0.44 (95% confidence interval: 0.39, 0.49) was observed overall, due in large part to a low cigarette smoking prevalence. Two widely used herbicides, metolachlor and pendimethalin (for low-exposed groups to four higher exposure categories: odds ratio (OR) = 1.0, 1.6, 1.2, 5.0; p(trend) = 0.0002; and OR = 1.0, 1.6, 2.1, 4.4; p(trend) = 0.003, respectively), and two widely used insecticides, chlorpyrifos and diazinon (OR = 1.0, 1.1, 1.7, 1.9; p(trend) = 0.03; and OR = 1.0, 1.6, 2.7, 3.7; p(trend) = 0.04, respectively), showed some evidence of exposure response for lung cancer. These excesses could not be explained by previously identified lung cancer risk factors. The usage levels in this cohort are considerably higher than those typically experienced by the general population. An excess risk among spouses directly exposed to pesticides could not be evaluated at this time.