Epidemiological data have not yet enabled physicians to look beyond age and race to identify men at increased risk for prostate cancer. We conducted a hospital-based case-control study of familial patterns of prostate cancer with self-reported data from a risk-factor questionnaire. There were 385 patients with histologically confirmed prostate cancer, and 385 race and age-matched (+/- 5 years) controls with other cancers. Family history, available for 378 patients and 383 controls, was positive for prostate cancer in 13.0% versus 5.7%, respectively. The difference was significant at p = 0.01. The over-all age-adjusted risk estimate for men with a first-degree relative with prostate cancer was significantly elevated (odds ratio of 2.41), as were the individual risk estimates for having a father or brother with prostate cancer (odds ratio of 2.24 and 2.66). Having a second-degree relative (grandfather or uncle) with prostate cancer also conferred elevated but not statistically significant risk. These data accord well with the few previously published case-control studies of familiarity of prostate cancer. On the basis of these findings, one should consider recommending participation in early detection programs for prostate cancer in a man whose father or brother has had the disease.