Epidemiologic evidence on the relation between sunlight and cancer is reviewed. Strong evidence implicates sunlight as a cause of skin cancer, although, for melanoma and basal cell carcinoma, the relationship is complex. Both types of cancer are associated more strongly with nonoccupational exposure than with occupational exposure, and the pattern and amount of exposure each appear to be important. Squamous cell carcinoma appears to be related more strongly to total (i.e., both occupational and nonoccupational) exposure to the sun. The evidence that sunlight causes melanoma of the eye is weak. It shows no latitude gradient and the results of case-control studies are conflicting. There is inadequate evidence to suggest that sunlight does or does not cause any other type of cancer.