The carcinogenicity of industrial formaldehyde (50000) exposures was investigated among workers producing shirts from formaldehyde treated cloth. Formaldehyde concentrations were measured at three manufacturing facilities. Deaths reported to an employee death benefit insurance fund were examined. Frequency distributions of deaths according to sex, race, facility, age at death, and duration of employment were studied and proportional mortality ratios (PMRs) calculated. Employees were predominantly white, 75 percent female, and 67 percent over 55 years of age at the time of the study. Air concentrations of formaldehyde ranged from 0.1 to 1.0 part per million. No deaths due to nasal cancer were observed. Statistically significant excesses in mortality were observed for cancers of the buccal cavity, PMR 750, biliary passages and liver, PMR 400, and other lymphatic and hematopoietic sites, PMR 400. Statistically nonsignificant excesses in mortality were evident in many of the malignancy categories. There was a slight excess in mortality due to diseases of the circulatory system, PMR 104, and respiratory system, PMR 118. Marked deficits in mortality were observed for diseases of the digestive system, 1 being observed versus 9.9 expected. Statistically significant excesses were present in the category of workers with greater than 10 years of latency and duration of exposure for all malignancies, PMR 137, buccal cavity cancer, PMR 952, malignancies of the biliary passages and liver, PMR 467, malignancies of all lymphatic and hematopoietic sites, PMR 283, and for malignancies of other lymphatic and hematopoietic sites, PMR 761. There was also a statistically significant excess in mortality due to malignancies of the digestive organs and peritoneum, PMR 194, in the less than 10 years of latency and exposure category. The authors conclude that observed excesses in cancer mortality could be related to occupational formaldehyde exposure.