Mortality through 1988 was studied for 5,932 male employees who worked between January 1, 1946 and December 31, 1967 at a New Jersey plastics manufacturing and research and development facility. The cohort was followed for an average of 32 years and included 1,859 deaths. Potential exposures included asbestos, formaldehyde, and polyvinyl chloride (PVC). Mortality rates for the cohort were compared to both U.S. and state mortality rates, and analyses were also performed by lagging duration of employment. Based on U.S. rates, mortality among hourly males (n = 3,853) from all cancers was similar to expected [standardized mortality ratio (SMR), 102; 95% confidence interval (CI), 92-114]. Excess mortality among hourly workers was seen for pancreatic cancer (SMR, 146; 95% CI, 95-216) and "malignancies of other parts of the respiratory system" (SMR, 373; 95% CI, 121-870). The latter excess was due entirely to five deaths from pleural mesothelioma. There were no deaths identified due to nasal cavity or nasopharyngeal cancers, or angiosarcoma of the liver. Mortality from leukemia among research and development workers (n = 1,421) was significantly elevated (SMR, 265; 95% CI, 115-524) and related to assignment to process development. This study verifies the excess of pancreatic cancer among workers at the facility seen in earlier studies and observes excesses of mesothelioma due to asbestos exposure and leukemia in process development workers.