Interviews were performed on 83 patients with malignant melanoma, being 74% of all new NHS patients over a 33 month period who were resident in a defined area of Nottingham, and on age and sex matched controls chosen from all outpatients and inpatients of the same hospitals with the same area of residence. Significantly increased risks of melanoma were found in subjects with 3 or more raised moles on the upper arms (relative risk = 17.0), in association with heavy freckling of the face and arms, and with a tendency to sunburn easily and tan poorly, these factors having independent effects. While no significant and consistent association with exposure to fluorescent light was seen, the observed risks were higher in subjects with greater exposure, and higher in association with exposure to undiffused than to diffused light. Cases had a significantly greater number of hours' exposure to undiffused light than did controls. The associations with fluorescent light exposure were stronger when based on interview data than on a subsequent postal questionnaire. Twenty-one cases and 11 controls reported exposure to unusual occupational lighting sources which may have had an ultraviolet component; these included various intense lighting sources and lamps used in printing and dyeline copying.