The objectives of this study were to investigate the effect of family history on prostate cancer risk, to estimate the incidence of hereditary prostate cancer in southern Sweden and to assess the reliability of self-reported family history of prostate cancer. The study included consecutive prostate cancer patients and age-matched control subjects from a geographically defined population. The controls consisted of 1 male patient with malignant melanoma or non-Hodgkin's lymphoma and 1 male from the community per prostate cancer case. Family history was assessed with questionnaires, and diagnoses of fathers and brothers of cases were validated by the Southern Swedish Regional Tumour Registry. Among fathers and brothers whose names and birth dates were available, 56 (92%) of the 61 reported prostate cancer diagnoses were verified. Fifteen per cent of 356 cases and 5.0% of 712 controls reported at least 1 case of prostate cancer among their brothers or fathers, giving a relative risk of 3.2 (95% confidence interval 2.1-5.1). The relative risk increased with decreasing age at diagnosis of the patient. Based on the pedigree, 3.1% of the 356 patients were classified as having hereditary prostate cancer. This proportion was significantly higher among patients diagnosed before the age of 60 years (7.1%) than among older patients (2.2%). We conclude that there is a substantially increased risk of prostate cancer for sons and brothers of prostate cancer patients. The risk increases with decreasing age at diagnosis of the patient as an effect of a higher prevalence of hereditary prostate among early onset cases. Furthermore, we found self-reported family history of prostate cancer to be a valid estimate of the true incidence of prostate cancer in fathers and brothers of men with prostate cancer.