Summary Relapse occurs in 50% of patients receiving radiation for clinical stage (C.S.) I and II nodal and extranodal non-Hodgkin's lymphoma (N.H.L.). Prior to the introduction of intensive chemotherapy those failing primary control with irradiation and most of those who relapsed died of their disease with a resultant overall mortality of 50%. An analysis of Princess Margaret Hospital results with radiation for C.S. I and II N.H.L. between January 1967 and December 1978 revealed that tumour bulk, age, stage and histology were of independent prognostic significance. It was possible to group patients using combinations of these attributes so that each group encompassed only patients with similar outcomes. Such prognostic groups were identified separately within the low grade and the intermediate plus high grade categories of the Working Formulation. Patients with a high probability of cure with radiation were so defined. Also those patients in whom chemotherapy would be optimal initial therapy were also defined. Such patients were in the intermediate plus high grade histology groups. Thirty percent of all patients with low grade histology lymphoma had an actuarial survival of 83%, and relapse-free rate of 63% at 10 years. By implication, approximately 20% of all patients with these histologies seen at the Princess Margaret Hospital for the same time period achieved prolonged relapse-free survival by localized therapy. This is at variance with the implications of staging from studies where laparotomy and multiple bone marrow biopsies have been used. Such aggressive staging procedures suggest truly localised disease in only 5–6% of patients with low grade lymphoma. A significant relationship between radiation dose and disease control was demonstrated only for patients with intermediate and high grade lymphoma of medium or large bulk. A minimum tumour dose of 30 Gy was required for optimal local control with radiation.