From July 1979 through March 1984 the Radiation Therapy Oncology Group conducted a randomized study comparing fast neutron radiotherapy versus mixed beam (neutron/photon) radiotherapy versus conventional radiotherapy for patients with non-small cell carcinoma of the lung. Patients were either medically or technically inoperable. One hundred two evaluable patients were placed on the study. The radiation doses were approximately 60 Gy-equivalent on each arm. Patients were stratified according to size of primary, histology, Karnofsky performance status, and age distribution. Overall local response rates as measured by serial radiographs were the same on the three arms, and an actuarial analysis showed no significant differences in either median or long-term survival. However, for the subgroup of patients exhibiting a complete or partial tumor response at 6 months there was a suggestion of improved 3-year survival on the two experimental arms (mixed beam, 37%; neutrons, 25%; photons, 12%). The p value for the difference between the mixed beam and photon curves is 0.14 (two-sided test). The incidence of major complications was higher on the neutron and mixed beam arms. These complications included four cases of myelitis which are analyzed in detail. The results are placed in the context of other published work on the use of neutrons in the treatment of lung cancer.