We have conducted a cohort mortality study on 689 patients with beryllium disease who were included in a case registry. An earlier mortality study on 421 of these patients was limited to males and resulted in a determination of a nonsignificant twofold lung cancer excess based on only seven lung cancer deaths. We have extended this earlier study by including females and by adding 13 years of follow-up. Comparison of the 689 beryllium disease patients with the U.S. population resulted in a lung cancer standardized mortality ratio (SMR) of 2.00 (95% confidence interval = 1.33-2.89) based on 28 observed lung cancer deaths. Adjustment for smoking did not change these results. All causes of mortality were also significantly elevated (SMR = 2.19), largely because of the very high rate of deaths due to pneumoconioses (primarily beryllium disease) (SMR = 34.23; 158 deaths). No other causes of death were significantly elevated. The excess of lung cancer was consistent for both sexes and did not appear to increase with duration of exposure to beryllium or with time elapsed since first exposure to this element. The case registry included those with acute beryllium disease, which resembles a chemical pneumonitis, and those with chronic beryllium disease, which resembles other pneumoconioses. The lung cancer excess was more pronounced among those with acute disease (SMR = 2.32) than among those with chronic disease (SMR = 1.57).