A case-control study was conducted in Los Angeles County, CA, of 75 male breast cancer cases aged 20-74 yr at diagnosis to investigate the role of a number of suspected risk factors. The study involved both interviews and laboratory measurements. Factors under study included fertility and marital history, obesity, alcohol and cigarette consumption, use of drugs known or suspected of causing gynecomastia, family history of breast cancer, history of radiation exposure to the upper body, sex chromatin analysis, serum levels of prolactin, testosterone, estrone, estradiol and sex-hormone-binding globulin, as well as urinary levels of estrone, estradiol, and estriol. Two patients versus no controls tested positive for sex chromatin and were excluded from further analyses. The only statistically significant risk factor identified was greater weight of the cases at age 30; a man who weighed 80 or more kg at age 30 had twice the risk of breast cancer of a man weighing less than 60 kg at that age. Serum estrone levels were positively, and sex-hormone-binding globulin levels were negatively, related to body weight, and we interpret the greater weight of the cases as suggesting that the underlying risk factor is an increased exposure to bioavailable estrogen. None of the differences observed between cases and controls for either the serum or urinary hormone levels was, however, statistically significant and there did not appear to be any large absolute excess of estrogens or deficit of testosterone in the cases. This apparent contradiction may be explained by the fact that there was little difference in weight between the cases and controls at the time of sampling.