PIP: This report is based on a study of 19 patients with breast cancer, 11 patients with atypical endometrial hyperplasia, and 7 women as controls. Single 24-hour urine collections were made for steroid determinations on Days 22 or 23 of the menstrual cycle. Among the patients with atypical endometrial hyperplasia, with or without breast cancer, a high incidence of involuntary sterility was noted. The range of testosterone excretion values with atypical endometrial hyperplasia w as significantly increased (p less than .01). In the breast cancer patients in whom the menstrual cycle was anovulatory and the endometrium obtained at the premenstrual period showed an atypical endometrial patte rn the testosterone excretion level was markedly increased (p less than . 01). During the follow-up period of 2-3 years, 5 of the 10 breast cancer patients with atypical endometrial patterns had developed distant metastases. Of the 9 breast cancer patients with ovulatory menstrual cycles, as shown by normal progestational endometrium, only 1 developed distant metastases. This patient had the highest urinary testosterone value in the group. In view of the high testosterone excretion levels, 7 cancer patients were offered ovariectomy. 4 refused but then underwent ovariectomy 4-8 months later for metastases. Of the 3 who accepted early ovariectomy, none have yet had metastases. The removed ovaries were shown to have polycystic ovarian disease with luteinized theca cells in the follicular cysts and marked interstitial cell hyperplasia. It was thought that androgens must play an important role in the development of breast cancer and endometrial hyperplasia. The in creased level of estrogens may have resulted from peripheral conversion. Ovarian interstitial tissue has been suggested as being a distinct gland of internal secretion that is principally concerned with the formation of androgens.