In a case-control study, we investigated 169 women aged 15-49 years with malignant melanoma notified to the Oxford and South Western cancer registries during the years 1971-1976, together with 507 matched controls. Data about medical, reproductive, drug and smoking histories were obtained both by reviewing general practitioner (GP) records and from the women themselves by postal questionnaires. There was no significant evidence of any overall increase in the risk of melanoma in oral contraceptive (OC) users (data from GP records-ever use vs never use, relative risk (RR) 1.34, 95% confidence limits 0.92-1.96; corresponding data from postal questionnaires-RR 1.13, limits 0.73-1.75). However, although not significant, the risk estimated from data in the postal questionnaires was higher in women who had used OCs for 5 years or more (use greater than or equal to 5 years vs never use, RR 1.57, limits 0.83-3.03). Previously demonstrated risk factors for melanoma, such as fair skin, blond or red hair and Celtic origin were found to be commoner in the cases than in the controls. Data from the Oxford/Family Planning Association contraceptive study were also examined. Unexpectedly there was a strong suggestion of a negative association between OC use and melanoma risk, but the analysis was based on only 12 women with the disease.