Two-hundred-thirty-two evaluable patients with metastatic sarcomas received a palliative experimental treatment program consisting of Adriamycin, 60 mg/m2, cyclophosphamide, 600 mg/m2, and methotrexate, 25 mg/m2 intravenously every three weeks for two courses. This program was followed by a randomization to maintenance therapy consisting of either the same regimen (ACM), an alternative regimen ADV (actinomycin-D, 1 mg/m2, DTIC, 250 mg/m2, and vincristine, 1.4 mg/m2 intravenously weekly), or alternating ACM and ADV. Twenty percent of patients had complete (6%) or partial (14%) responses, and an additional 41% of patients were stable throughout the remission and maintenance regimens. Response rates were similar regardless of the degree of toxicity observed during induction therapy. Although improved quality of responses after maintenance therapy was more frequent on ACM-ADV (19%) than on ADV (3%) or ACM (10%) maintenance therapy, the durations of response were similar for the three regimens (25 weeks, 19.7 weeks, and 19.6 weeks, respectively). Complete responses lasted a median of 44.5 weeks versus 19 weeks for partial responses. Although the length of survival depended upon the quality of response ultimately achieved (16 months for complete responders, 13.6 for partial responders, 11.6 months for patients with stable disease versus 4.0 months with progressive disease), median survival was similar for patients on each of the three randomized maintenance regimens (12 months for ACM, 13 months for aDV, and ten months for ACM-ADV). Toxicity was significantly worse on either of the regimens containing ADV, particularly peripheral neuropathy and gastrointestinal toxicity. Therefore, the ACM regimen for remission induction and remission maintenance is the most desirable of the treatment programs tested.