A case control study of 244 prostate cancer patients and 240 of their female partners was performed. The study sample was generated by the following methods: 1) an advertisement was placed in two newspapers with province-wide distribution 2) all Nova Scotia urologists were asked to identify patients with a positive family history and 3) consecutive patients treated for prostate cancer in an ambulatory setting were contacted by telephone. Seven men and 27 women did not know their family history. Patients with a positive family history (43/237, 18%) became case group. Wives who had family members affected by prostate cancer (16/213, 7.5%) comprised the control group. Both cases and controls completed a standardized questionnaire by mail or telephone. The heredity variant of prostate cancer accounted for 5 %, the familial 12% and the sporadic 83% of cases studied and 0%, 9% and 91%, respectively of the control group. The odds ratio calculated for first-and second-degree relatives was 2.73 (95% confidence interval, 1.49-5.01). The mean age at diagnosis of these cases was 68 years. There was not a statistically significant difference between the familial (69 years) and hereditary (67 years) subgroups' mean age at diagnosis (p=0.4). Most men with prostate cancer in the Nova Scotia population sample were affected by the sporadic variant. Small but important groups were affected by familial or hereditary types. Identification of these families will have important implications for screening and gene mapping studies.