Both red bone marrow and male breast doses with associated uncertainty have been reconstructed for a 1,982-person subset of a cohort of 114, 270 military personnel (referred to as “atomic veterans”) who participated in U.S. atmospheric nuclear weapons testing from 1945 to 1962. The methods used to calculate these doses and corresponding uncertainty have been reported in detail by Till et al. in an earlier publication. In this current article we report the final results of those calculations. These doses are being used in a case-cohort design epidemiological investigation of leukemia and male breast cancer. This cohort of atomic veterans is one component in a broader-scope study of approximately one million U.S. persons designed to investigate risk from chronic low-dose radiation exposure. Doses to the atomic veterans in this sub-cohort were relatively low, with approximately two-thirds receiving red bone marrow doses <5 mGy and only four individuals receiving a red bone marrow dose >50 mGy. The average red bone marrow dose for members of the sub-cohort was 5.9 mGy. Doses to male breast were approximately 20% higher than red bone marrow doses. The uncertainty in the estimated doses was relatively low, considering relevant personnel dosimetry was available for only about 25% of the subjects, and most of the doses were reconstructed from film badges worn by co-workers or from the individual's military record and military unit activities. The average coefficient of variation for the individual dose estimates was approximately 0.5, comparable to the uncertainty in doses estimated for the Japanese A-bomb survivors. Although the reconstructed red bone marrow doses were about 36% lower on average than the conservative doses previously estimated by the military for compensation, the overall correlation was quite good, suggesting that the estimates of doses from external exposure by the military for all ∼115,000 cohort members could be adjusted appropriately and used in further epidemiological analyses.