The role of radiotherapy and/or surgery in the local treatment of Ewing's sarcoma has still to be determined. The outcome of Ewing's sarcoma may differ according to its location and a selection bias towards surgery limits the ability to compare methods of local treatment. We have carried out a retrospective review of 91 consecutive patients treated for non-metastatic Ewing's sarcoma of the femur. They received chemotherapy according to four different protocols. The primary lesion was treated by surgery alone (54 patients), surgery and radiotherapy (13) and radiotherapy alone (23). One was treated by chemotherapy alone. At a median follow-up of ten years, 48 patients (53%) remain free from disease, 39 (43%) have relapsed, two (2%) have died from chemotherapeutic toxicity and two (2%) have developed a radio-induced second tumour. The probability of survival without local recurrence was significantly (p = 0.01) higher in patients who were treated by surgery with or without radiotherapy (88%) than for patients who received radiotherapy alone (59%). The five- and ten-year overall survival rates were 64% and 57%, respectively. Patients who were treated by surgery, with or without radiotherapy, had a five- and ten-year overall survival of 64%. Patients who received only radiotherapy had a five- and ten-year survival of 57% and 44%, respectively. Our results indicate that in patients with Ewing's sarcoma of the femur, better local control is achieved by surgical treatment (with or without radiotherapy) compared with the use of radiotherapy alone. Further studies are needed to verify the impact of this strategy on overall survival.