Combined radiotherapy and intensive chemotherapy have led to improved prognosis in children with non-Hodgkin's lymphoma, but the toxic effects of these forms of treatment are additive and seriously limit the benefits of their use. In an effort to limit the adverse acute and late effects of treatment, we conducted a randomized, controlled trial to determine whether irradiation of primary sites of involvement could safely be omitted from the treatment of children with localized non-Hodgkin's lymphoma (Stages I and II) who have a favorable prognosis. In addition, the chemotherapy regimen was less intense and shorter (eight months) than usual. A total of 129 patients were randomly assigned to receive either chemotherapy (vincristine, cyclophosphamide, doxorubicin, prednisone, mercaptopurine, and methotrexate) with irradiation (27 Gy) of the involved field (combined chemotherapy and radiotherapy) or chemotherapy alone. Half the patients have been followed for more than 38 months (range, 13 to 68), and the projected disease-free survival (+/- SE) at four years among patients assigned to chemotherapy alone is 87.9 +/- 8.8 percent, as compared with 87.3 +/- 9.4 percent among patients assigned to combined therapy (P = 0.44). The majority of patients tolerated therapy very well, although the toxic effects of treatment (myelosuppression, mucositis, and infection) were significantly worse among patients who received the combination of chemotherapy and radiotherapy. We conclude that most children with localized non-Hodgkin's lymphoma can be cured by a chemotherapy regimen of reduced intensity and duration. Radiotherapy can be safely omitted from the therapy of such children without substantially jeopardizing their excellent chance of cure.