Between October 1976 and May 1984, 156 patients with locally advanced cervical cancer were entered into a Phase III trial with the participation of five institutions. Patients were randomly assigned to receive photons only (50 Gy in 25 fractions over 5 weeks plus intracavitary applications or external-beam boost) or mixed-beam radiotherapy (2 fractions a week of neutrons, 3 fractions a week of photons to a total RBE-adjusted dose of 50 Gy over 5 weeks plus intracavitary applications or external mixed-beam boost). Only patients with squamous carcinoma of FIGO Stages IIB, III, or IVA with negative para-aortic nodes on lymphangiogram were eligible. Ten patients were excluded from the analysis because of ineligibility or cancellation. Of the 146 patients analyzed, 80 were treated with mixed-beam radiotherapy and 66 with photons. Patients were grouped by stage and institution. The percentage of patients undergoing intracavitary applications was 50% on mixed beam and 75% on photons (p less than 0.01). Tumor clearance was 52% and 72% for mixed beam and photons, respectively (p less than 0.03). Local control at 2 years was 45% for mixed beam and 52% for photons. Median survivals were 1.9 years on mixed beam and 2.3 years on photons. Severe complications occurred in 19% and 11% in mixed beam and photons respectively (p less than 0.13). The inferior outcome with neutron therapy in this study may have resulted from the use of horizontal neutron beams of varying energy and penetration. A new randomized trial using high-energy hospital-based cyclotrons with gantry-mounted beam-delivery systems has recently been activated to evaluate more rigorously the role of fast-neutron therapy for advanced cervical cancer.