While it is recognized that relatives of melanoma patients are at increased risk for this disease, the source and extent of variation in melanoma risk between families of melanoma cases is unknown. Heterogeneity of familial melanoma risk was assessed among the families (comprising 7,666 first-degree relatives) of 1,149 cutaneous melanoma cases diagnosed in Queensland, Australia, between 1982 and 1987. The measure of familial melanoma risk was based on the number of cases of melanoma in the family in excess of those predicted from the age-, sex-, and birth cohort-specific cumulative incidences of melanoma among all relatives in the sample. Probands over-reported melanoma occurrence among their relatives, with a false positive reporting rate of 44.5% (216 false reports out of 485). Only medically verified cases among relatives were included in the analysis. There was statistically significant heterogeneity in family risk, with 53 (4.7%) of the total 1,116 unrelated families containing significantly more melanoma cases than expected considering the size of the family, and the age, sex, and birth cohort distributions of family members. In univariate analyses, members of the high-risk families were significantly more likely to have poor ability to tan, a propensity to sunburn, fair skin color, red hair, and many melanocytic nevi. When all variables were included simultaneously in a multiple logistic regression model, only the associations with tanning ability, skin color, and number of nevi remained significant. There were no significant differences overall between high-risk and other families in the sites and ages at diagnosis of melanoma, although melanomas on variably sun-exposed sites (trunk and legs) were diagnosed earlier in the high-risk families, independent of the stage at diagnosis.