Two papers describing the role of constitutional factors in the etiology of melanoma are in press. One describes the results of a case-control study conducted using 185 pairs of North American twins discordant for malignant melanoma. Blonde and especially red hair were strong and significant predictors of melanoma in DZ pairs (ORs = 3.9 and 14.9, respectively). Childhood freckling was linked to risk and was of marginal significance in MZ twins (OR = 3.6) as was earlier age at the time of worst sunburn (OR = 2.1). The strongest factor, as in other studies of melanoma, was the presence of large moles (>7 mm in diameter). The second is a case-control study conducted in Los Angeles County with the same goals. Representative melanoma cases among non-Hispanic white residents under age 65 occurring over a discrete period were compared to controls were matched to cases by age, sex, race and neighborhood of residence, and cases and controls were interviewed in their homes by the same interviewer. Skin color was measured using a Photovolt #670 reflectance meter. The number of all moles and those over 7 mm in diameter on the right forearm, the density of freckles, and a subjective estimate of all large moles on the body were recorded. The important pigmentary risk determinants were skin reflectance (OR = 2.0, 2.7 for medium and light skin tones respectively), hair color (OR = 2.5, 2.6, for blondes and redheads respectively), and freckles (OR = 1.7). Two standard indicators of pigment density (eye color and propensity to burn) showed no significant effect after adjustment for other factors. Both the size and the number of moles conferred risk (OR = 2.7 for any large forearm mole and OR = 3.3 for >4 large body moles). These studies will serve as the basis for studying the interrelationships between the most important known environmental cause of melanoma, sunlight exposure, and these constitutional factors.