One hundred and ninety-five patients with metastatic malignant melanoma were randomized to receive either cyclophosphamide 600 mg/m2 IV plus DTIC 600 mg/m2 IV day 1 (CD); or cyclophosphamide 400 mg/m2 IV, DTIC 400 mg/m2 IV, and piperazinedione 4 mg/m2 IV on day 1 (PCD). Therapy was repeated every 21 days. Patient groups were similar regarding pretreatment performance status, evaluability, and site of metastases. The overall response rate was low, 11% on CD and 12% on PCD. Paradoxically, patients with visceral disease responded at least as frequently as patients with skin and lymph node metastases only (12% and 6% respectively for CD, and 15% and 5% for PCD). Survival was identical on each treatment program, with medians of six months. The major dose-limiting toxicity was myelosuppression, which was similar on each treatment program. We conclude that the addition of piperazinedione to cyclophosphamide plus DTIC does not improve the response rate in patients with metastatic malignant melanoma. Both of the treatment programs (CD and PCD) utilizing one-day DTIC produced response rates slightly (but not meaningfully) lower than those previously obtained with cyclophosphamide plus five-day DTIC.