The German Society of Pediatric Oncology in 1981 initiated the Cooperative Ewing's Sarcoma Study (CESS 81) using a four-drug combination of chemotherapy prior to definitive local control with surgery and/or radiation. From January 1, 1981 until February 28, 1985, 93 patients were registered at the trial office from 54 participating institutions in West Germany, Austria, Switzerland, and the Netherlands. On February 1, 1987, 54 of 93 patients were disease-free. Using the Kaplan-Meier life table analysis, the estimated disease-free survival (DFS) rate was 60% at 36 months and 55% at 69 months. The median period of observation was 29 months, ranging from 22 months to 69 months. Twenty-one of 93 patients (23%) had local failure, 18 of 93 patients (19%) developed systemic metastases. The local failure rate was particularly high in patients treated with radiation and was reduced when radiation planning was centralized within the study based upon the extent of disease at diagnosis. Cox regression analysis of prognostic factors showed that tumor volume was a significant factor influencing prognosis. The estimated 3-year DFS rate was 80% for patients with small tumors (volume less than 100 ml) compared to 31% for patients with large tumors (volume greater than or equal to 100 ml). In patients who had surgery for local control, the histologic response to chemotherapy was analyzed on the surgical specimen and had a strong influence on survival: 79% DFS at 3 years for patients with less than 10% viable tumor (good responders) compared to 31% DFS for patients with more than 10% viable tumor (poor responders). Tumor load and responsiveness to chemotherapy are the two major factors influencing prognosis in patients with primary Ewing's sarcoma of bone.