A total of 270 patients with metastatic malignant melanoma were entered into a randomized chemotherapy study conducted by the Central Oncology Group (COG) over a period of 2 years (COG protocol No. 7130). The study utilized DTIC, CCNU, BCNU, vincristine, and hydroxyurea. The results of therapy with DTIC alone were compared with three combinations: (a) DTIC, CCNU, and vincristine; (b) DTIC, BCNU, and vincristine: and (c) DTIC, BCNU, and hydroxyurea. There were 243 evaluable patients out of 270. The response rate was 17.3% (42 of 243 patients) for evaluable patients and 15.5% (42 of 270 patients) for all patients entered in the study. The results showed no statistically significant difference in the response rates among the four treatment arms. However, several significant points were observed: (a) A 13% response rate was obtained using a combination of agents that included DTIC in patients who had previously shown no response to DTIC used as a single agent. (b) There was a significant difference in survival time when comparing responders to those with no change and to those with progression. (c) Toxicity was noted to be greater in responders than in nonresponders. (d) Two of the four treatment arms were considered most advantageous due to the ease of administration. These treatment arms were DTIC, BCNU, and vincristine and DTIC, CCNU, and vincristine administered in 5-day courses every 6 weeks. (e) The percentage of response and length of survival were significantly greater in patients without brain or liver metastasis. (f) In comparing men to women there was no statistically significant difference in response in rates or durations of response. (g) There was no statistically significant difference in survival when comparing site of primary lesion.