The utility of a population based, corporate wide mortality surveillance system was evaluated after a 10 year observation period of one of the company's divisions. The subject population, 2219 white male, long term employees from Union Carbide Corporation's carbon based electrode and specialty products operations, was followed up for mortality from 1974 to 1983. External comparisons with the United States male population were supplemented with internal comparisons among subgroups of the study population, defined by broad job categories and time related variables, adjusting for important correlates of the healthy worker effect. Significant deficits of deaths were observed for all causes and the major non-cancer causes of death. The numbers of deaths due to malignant neoplasms and respiratory cancer were less than, but not statistically different from, expected. There was a non-significant excess of deaths from lymphopoietic cancer, occurring predominantly among salaried employees. When specific locations were examined, operations with potential exposure to coal tar products exhibited a mortality pattern similar to that of the total cohort. The risk for lung cancer was significantly raised (five observed, 1.4 expected) in one small, but older, location which did not involve coal tar products during the period of employment of these individuals, but which historically used asbestos materials for several unique applications. Although these findings are limited by small numbers and a short observation period, the population based surveillance strategy has provided valuable information regarding the mortality experience of the population, directions for future research, and the allocation of epidemiological resources.