Potential prognostic factors for survival time, 5-year survival rate, recurrence-free time, and 5-year recurrence-free rate were explored for 262 patients with malignant melanoma of the skin, Stage I, using the long-rank test, proportional hazards (Cox) regression, and linear logistic regression. Breslow's thickness was the most important variable in each analysis. A thinner Breslow's thickness was associated with a better prognosis. After adjusting for Breslow's thickness, sex was the next most important variable in each analysis. Women had a better prognosis than men. After adjusting for Breslow's thickness and sex, there were no additional significant prognostic factors for survival time; Clark's level was significant for 5-year survival rate; the number of mitoses, evidence of regression, and presence of vascular invasion were significant for recurrence-free time; and the presence of vascular invasion and number of mitoses were significant for 5-year recurrence-free rate. It is recommended that in future prospective studies of adjuvant therapy of Stage I melanoma, patients be stratified into treatment groups on the basis of Breslow's thickness and sex.