The histologic material from the original lesion of 113 patients suffering from Clinical Stage I and Clinical Stage II malignant melanomas were reviewed in an attempt to correlate depth of invasion with prognosis and to test whether or not current methods of microstaging were a valid guide to therapy. Both the microstaging methods of Clark and Breslow were eminently successful in indicating the prognosis of the patient, hut were of no greater utility than other histologic parameters such as mitotic activity, perineural, vascular, or lymphatic invasion, or the presence of superficial ulceration. The value of elective node dissection for the treatment of primary malignant melanoma is uncertain. Deterents against empiric dissection are its low yield of occult metastases, its inability lo prevent recurrence in a substantial portion of patients independent of the status of the nodes, and the difficulty in demonstrating an improved survival rate when (his method of therapy is employed. Neither Clark's nor Breslow's methods of microstaging appeared to yield adequate criteria by which to choose node dissection. A simplified method of microstaging is proposed utilizing well defined anatomic levels, and yields statistically valid criteria for the performance of elective node dissection. Using this method, elective node dissection is advocated solely for those tumors that clearly invade the reticular dermis. Patients with tumors restricted to the papillary dermis do not appear lo be benefited by empiric node dissection.