In 1984, based on epidemiological data on cohorts of coke oven workers, USEPA estimated a unit risk for lung cancer associated with continuous exposure from birth to 1 microgram/m3 of coke oven emissions, of 6.2 x 10(-4). This risk assessment was based on information on the cohorts available through 1966. Follow-up of these cohorts has now been extended to 1982 and, moreover, individual job histories, which were not available in 1984, have been constructed. In this study, lung cancer mortality in these cohorts of coke oven workers with extended follow-up was analyzed using standard techniques of survival analysis and a new approach based on the two stage clonal expansion model of carcinogenesis. The latter approach allows the explicit consideration of detailed patterns of exposure of each individual in the cohort. The analyses used the extended follow-up data through 1982 and the detailed job histories now available. Based on these analyses, the best estimate of unit risk is 1.5 x 10(-4) with 95% confidence interval = 1.2 x 10(-4)-1.8 x 10(-4).