One hundred eight-five patients with acral melanoma treated since 1972 were reviewed. These included 53 subungual lesions, 123 plantar lesions, and 9 palmar lesions. Eighty percent presented with stage I disease. Mean age was 57 years. Males outnumbered females 1.1:1. Seventeen percent (17%) were blacks. Actuarial 10-year survival was 58% for stage I patients and 35% for stage II patients. Univariate Cox regression analyses identified 5 prognostic variables affecting survival: stage at diagnosis (P less than 0.001), race (P less than 0.001), ulceration (P = 0.012), Clark's level (P = 0.014), and thickness of the primary lesion (P = 0.013). Factors unrelated to survival included sex of the patient, site (volar vs. subungual), histology, and treatment with amputation. Multivariate analysis for patients with stage I lesions identified race (P = 0.001) and ulceration (P = 0.018) as significant variables, with thickness approaching significance (P = 0.094). In an additional series of 71 patients with melanomas arising from extremity sites near the junction of glabrous and non-glabrous skin, survival was significantly poorer for those arising from glabrous skin (P = 0.024), and reflects a higher incidence of metastatic disease at diagnosis. Specific active immunotherapy was the principal adjuvant used for these patients, and survival was comparable to that reported with regional perfusion therapy. Acral melanoma a) has a strong racial predilection, b) carries a grave prognosis, and c) arises from glabrous skin. It is a clinical entity distinct from other extremity melanomas. Surgical management with either wide excision or amputation is appropriate for the primary lesion.