From 1981 up to February 1985, a total of 93 protocol patients entered the study CESS 81. The protocol recommended an initial 18-week period of polychemotherapy (VACA) followed by local therapy and two additional cycles of chemotherapy. Local therapy consisted either of radical surgery or of incomplete resection plus postoperative irradiation with 36 Gy or of radiotherapy alone (46 and 60 Gy). Centrally located lesions were always irradiated with 60 Gy. This article summarizes the data after 5 years. Data of 93 patients were analysed in October 1986 after a median follow-up of 37 months. The projected 5-year survival is 50%. The relapse rate was 42%, distant relapses occurred in 19%, local (plus distant) relapses in 23%. Most of the relapses occurred during the first 3 years of observation. Failure rate was high in patients undergoing irradiation alone (44%). Initial tumour mass (>100 ml) and histopathologic response to initial chemotherapy were identified as major prognostic factors. Tumour site and type of local therapy were not significant if patients were categorised by tumour volume. In small lesions, surgery and radiotherapy were equally effective. In large lesions >100 ml volume, a trend towards a better prognosis in surgically treated patients was observed. The results of CESS 81 emphasize the importance of permanent local control in Ewing's sarcoma even in the presence of systemic control by an effective multi-drug chemotherapy.