Between 1978 and 1984, the Northern California Oncology Group (NCOG) conducted a randomized trial to study the efficacy of combined radiotherapy (RT) and chemotherapy (CT) for stage III or IV inoperable head and neck cancer. One hundred four patients were randomized to receive: (1) RT alone, or (2) RT plus CT. RT consisted of 7,000 cGy to the involved areas and 5,000 cGy to uninvolved neck at 180 cGy/fraction, five fractions/wk. CT consisted of bleomycin, 5 U intravenously (IV), twice weekly during RT, followed by bleomycin, 15 U IV, and methotrexate, 25 mg/m2 IV weekly for 16 weeks after completion of RT. Fifty-one patients in the RT alone group and 45 in the combined treatment group were evaluable. The local-regional complete response (CR) rate was 45% v 67% (P = .056); the 2-year local-regional control rate, including salvage surgery, was 26% v 64% (P = .001); and the incidence of distant metastasis was 24% v 38% (P greater than .25), for the RT alone and RT plus CT groups, respectively. The relapse-free survival curves were significantly different (P = .041), favoring the combined treatment. However, the survival curves were not significantly different (P = .16). Patient compliance to maintenance CT was poor. Bleomycin significantly increased the acute radiation mucositis, although the difference in late normal tissue toxicity was not statistically significant. Thus, bleomycin and concurrent RT produced a more favorable CR rate, local-regional control rate, and relapse-free survival, but the difference in survival was not statistically significant.