We report here two randomized prospective clinical trials of adjuvant treatment in the management of primary malignant melanoma of Clark's level III, IV or V. All patients had curative resection of the primary tumor. In the first trial, 117 patients were randomized between control (surgery alone) systemic chemotherapy and intraarterial chemotherapy. Intraarterial chemotherapy consisted of DTIC 80 mg/m2 + 8 days prior to surgery. Systemic chemotherapy consisted of courses of vinblastine (6 mg/m2), thiotepa (6 mg/m2), rufocromycine (60 microgram/m2), methotrexate (15 mg/m2) on day 1, and procarbazine (30 mg/m12 X 7 days. Courses were repeated every 2 weeks X 6, then every 4 weeks X 15. Twenty-two of 55 patients relapsed in the control group versus 22 of 67 in the chemotherapy group (NS). For male patients, the difference in disease-free survival was significant (P less than 0.005, log rank test), though not in women. In the second trial, 352 patients were entered from July, 1976. Men were randomized between chemotherapy and chemoimmunotherapy. Women were randomized between surgery alone and chemoimmunotherapy. Chemotherapy was identical, except for the addition of DTIC (300 mg/m2) for each course. Immunotherapy consisted of BCG every 4 weeks and C. parvum every week. Immunotherapy seemed to be of no additional benefit.