The evidence relating cutaneous melanoma in humans to sun exposure was extensively reviewed by the International Association for Research on Cancer (IARC) in 1992. Since then, 10 further epidemiological studies have been published. These are systematically reviewed, in a manner consistent with the IARC Monograph, to provide a definitive update on melanoma epidemiology. As the development of melanoma by sun or ultraviolet exposure is not amenable to experimental testing in humans, and as there is no clearly analogous animal model, the evidence for this association is dependent on epidemiological studies, now numbering 39, involving more than 10,000 subjects. A review of the results from all these studies shows that there is strong evidence for increased risk related to intermittent sun exposure, whereas the relationships with occupation and with total sun exposure are much more variable. There is a strong and consistent association found with a history of sunburn. These results and their implications are discussed.