Fifty-three evaluable patients with disseminated ovarian carcinoma (FIGO III or IV) not treated with prior chemotherapy were randomized to receive either combination chemotherapy consisting of cis-platinum 40 mg/m2 IV on day 1, adriamycin 40 mg/m2 IV on day 1, and hexamethylmelamine 150 mg/m2 PO on days 2-10 up to a maximum of 200 mg on a 4-weekly cycle, or moderate-dose cyclophosphamide alone 40 mg/kg given IV intermittently every 3 weeks. Entry was from 1. 11. 1978 until 30. 4. 1981 (last follow-up 31. 10. 1981). Pretreatment characteristics in both groups of patients, regarding median age at diagnosis, median time from diagnosis to chemotherapy, FIGO stage, histology, differentiation grade, type of surgery, residual disease, previous radiotherapy, and median performance status, were comparable. Objective responses were seen in 18 of 27 (66%) of patients receiving cyclophosphamide alone (range 5--32+ months) and in 10 of 26 (38%) of patients treated with the combination (range 3--30+ months), this difference being statistically significant (chi 2 = 4.228; P less than 0.05). The median duration of objective response (11 vs 10 months) and the median survival (12 vs 11 months) were greater in the cyclophosphamide group, but these differences were not statistically significant. The toxicity of the combination was more severe. It is concluded that there is no therapeutic advantage for this combination schedule over the alkylating agent used alone.