Since the early 1970s, a number of case reports have suggested that levodopa therapy for Parkinson's disease increases the risk of cutaneous malignant melanoma. As yet, no formal epidemiological study has been conducted to verify this hypothesis. To elucidate the relationship between levodopa and the risk of cutaneous malignant melanoma, a systematic literature search using computerized bibliographic databases was done. This review presents the case history evidence for and against the hypothesis of a causal association, and explores possible epidemiological, genetic, social, biochemical and toxicological factors that may increase the risk of melanoma in Parkinson's disease patients. All the case reports in the literature were considered. We concluded that (1) there is no epidemiological or experimental evidence of a causal role of levodopa in increasing the risk of melanoma incidence or progression; (2) there is good evidence of an excess risk of melanoma in patients with Parkinson's disease; (3) there is good evidence of a protective effect of tobacco smoking on the risk for Parkinson's disease; (4) there is good evidence of positive correlation between social class and melanoma risk; (5) the relationship between the risk of Parkinson's disease and the risk of melanoma may be due to a common genetic profile or it can be attributed to a confounding role of social class, associated with both melanoma and Parkinson's disease possibly through an inverse relationship with tobacco smoking.