To investigate the relationship between occupation and lung cancer, a case-control study was performed in the province of Trieste, Italy, where metallurgical and mechanical industries, dock activities and shipbuilding and ship repairing are predominant. Through the local Cancer Registry, pathology records of 938 men who died of primary lung cancer (ICD 162) in a five-year period were examined. Residential, smoking and occupational histories were obtained from interviews of next of kin of 756 cases and 756 age-matched male controls (+/- 2 years). Occupational exposures to lung carcinogens were assessed according to a job-title based approach, identifying industries/occupations with well-recognized lung carcinogen exposures (list A) and industries/occupations with suspected lung carcinogen exposures (list B). Exposure to asbestos was classified as absent, possible or definite. After adjustment for cigarette smoking (four levels) and residence (three levels), a significant association was found between lung cancer and occupations in list A (RR = 2.28, 95% CI = 1.70-3.07) and in list B (RR = 1.33, 95% CI = 1.04-1.71). A significant excess risk was found for workers with definite exposure to asbestos when compared to those with no exposure to lung carcinogens (RR = 1.99, 95% CI = 1.43-2.76). A very high relative risk was observed among heavy smokers with definite exposure to asbestos (RR = 42.8). A stratified analysis showed that the combined effect of asbestos and smoking was compatible with that expected under a multiplicative model. The overall attributable risk in the population (ARp) for cigarette smoking was found to be 87.6%. The ARp fraction for occupations with well-established exposures to lung carcinogens (list A) was 16.2%. The ARp fraction increased to 25.5% (85% CI = 1.4-34.6) when occupations with suspected exposure to lung carcinogens (list B) were included. The ARp fraction for possible or definite exposure to asbestos was 20.1% (95% CI = 11.6-28.6).