The proportion of lung cancer related to occupational exposure to respiratory carcinogens and other workplace contaminants has been studied in a hospital-based case-referent study of 176 incident cases and 176 referents, recruited from the medical ward of two county hospitals in southern Norway. On the basis of job title and exposure information, three different approaches for the estimation of etiologic fraction were used, all based upon a trichotomous categorization of the exposure variable. Among the definitely exposed subjects, the etiologic fraction for lung cancer related to occupational exposure factors was from 22 to 35%, increasing to 37-47% when the "possibly exposed" categories were included. When asbestos was the only factor to be eliminated, the etiologic fraction estimate suggested a 23% reduction in the lung cancer incidence in the study population. More than 90% of the lung cancer cases would theoretically have been prevented if smoking had been completely removed. The relative merits of preventing one or several of the study factors simultaneously is discussed.