After stratification according to diameter of the largest residual tumor, 120 previously untreated ovarian cancer patients were randomized to receive adriamycin and cyclophosphamide in combination with hexamethylmelamine (HAC) or cis-dichlorodiamineplatinum (PAC). The surgical response rates were 66% to HAC and 70% to PAC, with median times to progression of 14 and 22 months and median survival times of 23 and 24 months, respectively. In patients with residual tumor greater than 2 cm the surgical response rates to HAC and PAC were 56% and 63%, with complete response rates of 13% and 21%, respectively. In two of five complete responders to HAC there has still been no progression at 38 and 48 months, with a median response duration of 25 months. Only one of the nine complete responders to PAC has relapsed, at 33 months, while in the eight others response is maintained at follow-up times of 35-64 months. Myelosuppression was generally mild and similar in the two arms. No significant nonhematological toxicity was reported. It is concluded that at a median follow-up time of 36 months HAC is as effective as PAC in terms of response, duration of remission, and survival in previously untreated advanced ovarian cancer.