To assess the association between the incidence of cutaneous melanoma; intermittent, occupational and total sun exposure; and history of sunburn at different ages, we conducted a systematic review using results of all published case-control studies which have assessed incident melanoma, sun exposure and sunburn. Twenty-nine studies contributed data on sun exposure and 21 on sunburn. Overall, there was a significant positive association (odds ratio [OR] = 1.71) for intermittent exposure, a significantly reduced risk for heavy occupational exposure (OR = 0.86) and a small, marginally significant excess risk for total exposure (OR = 1.18). There was a significantly increased risk with sunburn at all ages or in adult life (OR = 1.91) and similarly elevated relative risks for sunburn in adolescence (OR = 1.73) and in childhood (OR = 1.95). There was significant heterogeneity with all of these estimates except that of all ages or adult sunburn. These results show the specificity of the positive association between melanoma risk and intermittent sun exposure, in contrast to a reduced risk with high levels of occupational exposure. The association with sunburn also is likely to reflect intermittent exposure; the results do not suggest any strong relationship to age at sunburn. These associations are similar to those reported for basal cell skin cancer but different from those reported for squamous cell cancer. The mechanisms by which intermittent exposure increases risk, while other patterns of exposure do not, remain to be elucidated.