BACKGROUND Familial melanoma patients tend to have an earlier age at first melanoma diagnosis, thinner lesions, a different histologic distribution, and a higher frequency of multiple primary melanomas than patients with nonfamilial melanoma. Previous examination of a large melanoma kindred from Texas suggested that although cutaneous malignant melanoma (CMM) was transmitted in an autosomal dominant fashion, there were sex differences in penetrance and disease expression. OBJECTIVE This study further evaluated the age at diagnosis, sex difference in penetrance and disease expression, and segregation of familial CMM. METHODS We evaluated the age at diagnosis and transmission of CMM in 23 U.S. white families with CMM/dysplastic nevi who had been followed 5-17 years. We estimated the median and mean ages at diagnosis of invasive melanoma for all individuals, for men and women separately, and by generation. Using the computer program BMDP1L, we also estimated the cumulative probability of an offspring developing CMM as a function of age according to whether the mother or father had melanoma. In addition, we used a life-table approach to estimate the probability that offspring of CMM parents were affected with CMM (penetrance). RESULTS The median age at diagnosis in the 23 kindreds (n = 106) was 33 years, substantially less than that of patients with sporadic melanomas in the U.S. white population. For females, the median age at diagnosis was 29 years; for males, it was 36 years. Nine percent of the case patients developed CMM before age 20 compared with 2% in the general population. There was little difference in the transmission pattern of melanoma between males and females in the 23 families, although sons of CMM parents had a higher risk of CMM than daughters of CMM parents. This difference was, however, based on small numbers and was not statistically significant. The penetrance estimates for CMM were high. They rose rapidly from 6% at age 18 to 85% by age 48. The median age at diagnosis of invasive melanoma decreased dramatically in successive generations; the reduction was 11-16 years per generation and was statistically significant (P < .0001). CONCLUSIONS Although this reduction in median age at diagnosis may result partly from increased surveillance in hereditary melanoma families and the fact that individuals in the younger generations have not yet reached the highest at-risk ages for developing melanoma, the possibility of changes in melanoma risk and risk factors, genetic and/or environmental, across generations should be considered.