Between June 1981 and December 1985, 296 evaluable patients with carcinoma of the cervix (stages IIB, III, or IVA) were randomized to radiation therapy and either hydroxyurea (139 patients) or misonidazole (157 patients). All patients had undergone clinical, radiographic, and surgical staging. Patients with metastasis to periaortic nodes were ineligible for study. Patients received external radiation therapy to the pelvis and either one or two intracavitary applications. Hydroxyurea was given in a dose of 80 mg/kg each Monday and Thursday during external radiation therapy. Misonidazole was given in a dose of 1 gm/m2 in the same schedule, not to exceed 12 gm/m2. Of the evaluable patients, 60.8% had stage IIB disease and 33.8% had stage IIIB disease. Negative pelvic lymph nodes were found in 79.2% of the patients. Median age was 49 years (first and third quartiles 40 and 60, respectively). There were 51 patients who had severe and 15 patients who had life-threatening adverse effects (including two treatment-related deaths). As of February 1987 half the patients have either failed or been followed-up for at least 43 months. The group treated with hydroxyurea had a longer progression-free interval, bordering on statistical significance, than those treated with misonidazole (p = 0.08). The median progression-free interval for all patients randomized to hydroxyurea is 42.9 months and for misonidazole it is 40.4 months. The median progression-free interval for patients with stage III and IV disease who received hydroxyurea has not been reached and for the misonidazole group it was 10.1 months. There have been 120 recurrences, 51 (36.7%) in the hydroxyurea group and 69 (43.9%) in the misonidazole group; 51.7% of the recurrences have been limited to the pelvis or vagina. Failure limited to the pelvis occurred in 18.0% of patients receiving hydroxyurea and 23.6% of patients receiving misonidazole. There were 108 deaths, 47 (33.8%) in the hydroxyurea group and 61 (38.9%) in the misonidazole group; survival does not differ statistically between the two regimens at this point in follow-up (p = 0.25). Hydroxyurea has more short-term gastrointestinal and marrow toxicity, but is free of long-term neurotoxicity. Preliminary analyses indicate that there is no role for radiation therapy with misonidazole in cervical carcinoma.