A study of dose/response relationships for cancer mortality and morbidity in asbestos cement workers was conducted. The cohort consisted of 1929 males employed at a Swedish asbestos cement factory for at least 3 months between 1907 and 1977. The comparisons consisted of 726 workers from the same county who had no known asbestos exposure. Dose estimates for each asbestos (1332214) exposed worker were determined from company industrial hygiene monitoring data obtained between 1956 and 1977, use of engineering controls, and work histories. Data on the mineral fiber content of lungs from 76 deceased asbestos workers were reviewed. The vital status of the subjects was determined as of December 31, 1986. Cancer mortality and morbidity of the cohort were compared with that of the general population of Sweden. Mortality in the cohort from all cancers, respiratory cancer, nonmalignant respiratory disease, gastrointestinal cancer, and colorectal cancer was significantly increased, having standardized mortality ratios (SMRs) of 1.17, 1.98, 1.73, 1.46, and 1.64, respectively. Mortality from stomach cancer was nonsignificantly increased, SMR 1.35. Overall cancer morbidity and morbidity from gastrointestinal cancer were significantly increased. The lung tissue data indicated nine cases of mesothelioma. Tissue samples from these subjects contained primarily chrysotile (12001295) and crocidolite (12001284) at concentrations of 62x10(6) and 54x10(6) fibers per milligram, respectively. More than half of the respiratory cancer deaths in the cohort were from pleural mesothelioma, which were significantly related to length of employment. The other types of lung cancer showed no dose/response relationship. Mortality from gastrointestinal cancer was related to cumulative exposure. No dose response relationships were found for cancer morbidity or nonmalignant respiratory disease.