Within three weeks of definitive surgical intervention, 467 patients with histologically proved malignant glioma were randomized to receive one of four treatment regimens: semustine (MeCCNU), radiotherapy, carmustine (BCNU) plus radiotherapy, or semustine plus radiotherapy. We analyzed the data for the total randomized population and for the 358 patients in whom the initial protocol specifications were met (the valid study group). Observed toxicity included acceptable skin reactions secondary to radiotherapy and reversible leukopenia and thrombocytopenia due to chemotherapy. Radiotherapy used alone or in combination with a nitrosourea significantly improved survival in comparison with semustine alone. The group receiving carmustine plus radiotherapy had the best survival, but the difference in survival between the groups receiving carmustine plus radiotherapy and semustine plus radiotherapy was not statistically significant. The combination of carmustine plus radiotherapy produced a modest benefit in long-term (18-month) survival as compared with radiotherapy alone, although the difference between survival curves was not significiant at the 0.05 level. This study suggests that it is best to use radiotherapy in the post-surgical treatment of malignant glioma and to continue the search for an effective chemotherapeutic regimen to use in addition to radiotherapy.