For 1950-80, 194 ovarian cancer cases were ascertained among the 70,030 females of the Radiation Effects Research Foundation's Life-Span Study E-85 sample, and 106 autopsied cases with benign ovarian neoplasms were ascertained among all 3,046 autopsies performed in the same sample. On the basis of microscopic review, 66% of the cancer and 84% of the benign tumor cases were classified by histologic type. The age-adjusted ovarian cancer incidence rates showed a statistically significant increase with increased exposure dose, both in the entire exposed group (P less than .01) and in the microscopically reviewed subset (P less than .01). This dose response was only significant (P less than .01) in the latter half of the study period, 1965-80. The radiation effect was higher in the younger age group at the time of the bomb (ATB) for the specific attained age or was adjusted for attained age. In general, relative risk (greater than or equal to 100 rad vs. 0 rad) did not differ by attained age, except for the youngest age group, less than 20 years old ATB, where the relative risk tended to decrease with increased attained age, although cases were few in number and follow-up study was necessary. Estimated minimum latent period for radiation-induced ovarian cancer seemed to be 15-20 years. The proportion of autopsied cases with benign ovarian tumor increased with increasing exposure dose, both in the entire series of cases (P less than .05) and in the microscopically reviewed subset. Statistical significance, however, was not achieved in the latter group (P greater than .10). The distribution of histologic types of both cancer and benign tumor of the ovary did not vary significantly with radiation dose. The data are consistent with the hypothesis that radiation injury of the ovaries and secondary excess of gonadotropic hormones are important causative factors in the development of ovarian neoplasms.