To examine further the epidemiology of contralateral primary breast cancer, a case-control analysis, utilizing information available from the Connecticut Tumor Registry, was conducted. Recent cases of second primary breast cancer were compared to control women who had survived a first breast cancer but had not developed a second. Three hundred and thirty eight incident cases of contralateral breast cancer diagnosed between 1979 and 1982 were identified and compared with an equal number of randomly selected controls and 336 controls frequency matched to the cases on the basis of age at initial cancer diagnosis and the calendar time elapsing since that diagnosis. Risk of second primary breast cancer was found to be significantly elevated among women whose initial cancer was lobular carcinoma and during the first year following diagnosis of the initial primary. Additionally, for women initially treated with radiotherapy, risk of a contralateral primary increased for 10-14 years following treatment, after which it declined. Among young women, having never married was protective whereas the opposite was found among older women. These findings and the methods used are discussed in the context of the epidemiology of both contralateral and initial breast primaries.