A case-control study in Somerset and east Devon was undertaken to investigate possible risk factors for prostatic cancer. A total of 159 cases, diagnosed at Taunton. Yeovil and Exeter hospitals between May 1989 and May 1991, were identified prospectively and interviewed with a structured questionnaire. A total of 161 men diagnosed with benign prostatic hypertrophy and 164 non-urological hospital controls were given identical questionnaires. The questionnaire covered a wide range of factors identified from previous studies, but the central hypotheses for this study related to diet (fat and green vegetables), sexual activity and farming as an occupation. This study found no association between farming and risk of prostatic cancer (odds ratio = 0.74, 95% confidence interval 0.46-1.18), nor with sexual activity as measured by number of sexual partners (chi-squared test for trend P = 0.52). A history of sexually transmitted disease was not significantly associated with prostatic cancer, but the numbers involved were very small and the odds ratio of 2.06 (0.38-11.2) is consistent with the hypothesis. A range of questions aimed at eliciting dietary fat intake produced no significant associations, although meat consumption showed increasing risk with increasing consumption (test for trend P = 0.005). Increased consumption of leafy green vegetables was associated with lower risk, but not significantly so (test for trend P = 0.16). As expected with so many factors investigated, some statistically significant associations were found, although these can only be viewed as hypothesis generating in this context. These included apparent protective effects of circumcision and high fish consumption.