An analysis was conducted of 27,982 deaths among 106,020 persons employed at four Federal nuclear plants in Oak Ridge, Tennessee, between 1943 and 1985. The main objectives were to extend the evaluation of the health effects of employment in the nuclear industry in Oak Ridge to include most workers who were omitted from earlier studies, to compare the mortality experience of workers among the facilities, to address methodological problems that occur when individuals employed at more than one facility are included in the analysis, and to conduct dose-response analyses for those individuals with potential exposure to external radiation. All-cause mortality and all-cancer mortality were in close agreement with national rates. The only notable excesses occurred for white males for lung cancer [standardized mortality ratio (SMR) = 1.18, 1,849 deaths] and non-malignant respiratory disease (SMR = 1.12, 1,568 deaths). A more detailed analysis revealed substantial differences in death rates among workers at the Oak Ridge plants. Evaluation of internally adjusted log SMRs using Poisson regression showed that workers employed only at Tennessee Eastman Corporation or K-25 and at multiple facilities had higher death rates than similar workers employed only at X-10 or Y-12, and that the differences were primarily due to non-cancer causes. Analysis of selected cancer causes for white males indicated large differences among the workers at the different facilities for lung cancer, leukemia and other lymphatic cancer. Dose-response analyses for external penetrating radiation were limited to a subcohort of 28,347 white males employed at X-10 or Y-12. Their collective recorded dose equivalent was 376 Sv. There was a strong "healthy worker effect" in this subcohort-all-cause SMR = 0.80 (4,786 deaths) and all-cancer SMR = 0.87 (1,134 deaths). Variables included in the analyses were age, birth cohort, a measure of socioeconomic status, length of employment, internal radiation exposure potential and facility. For external radiation dose with a 10-year lag, the excess relative risk was 0.31 per Sv (95% CI = -0.16, 1.01) for all causes and 1.45 per Sv (95% CI = 0.15, 3.48) for all cancer. The estimated excess relative risk for leukemia was negative but imprecisely determined. A preliminary dose adjustment procedure was developed to compensate for missing dose but not other dosimetry errors. Results of the analyses using the adjusted doses suggest that the effect of missing dose is an upward bias in dose-response coefficients and test statistics.