Several recent studies (animal and human) have suggested an association between lung cancer and silica exposure. To test the hypothesis, we have studied death benefit records of 1,905 members of the Granite Cutters Union. A proportionate mortality analysis (PMR) was conducted, using U.S. deaths as a comparison population. Statistically (PMR) was conducted, using U.S. deaths as a comparison population. Statistically significant excesses were observed for death from nonmalignant respiratory significant excesses were observed for death from nonmalignant respiratory disease (largely silicosis) (183 obs, 43.7 exp) and for tuberculosis (largely silicotuberculosis) (262 obs, 19.3 exp). Other significant excesses were observed for bone cancer (6 obs, 1.9 exp) and arthritis (5 obs, 1.5 exp). A significant decrease was observed for leukemia (5 obs, 13.0 exp). For lung cancer a slight but nonsignificant excess was observed (97 obs, 81.1 exp, PMR = 1.19, 95% CI 0.97-1.46). A proportionate cancer mortality analysis (PCMR) showed similar results for lung cancer (PCMR = 1.09, 95% CI 0.89-1.33). Lung cancer mortality also failed to show any trend with either calendar time or duration of exposure. Although no significant excess of lung cancer was observed for the entire silica-exposed cohort, there was an indication that those who were silicotic had an excess risk of lung cancer, based on a review of contributing causes on the death certificate.