From 1968 to 1986, 192 patients from 0 to 17 years of age were enrolled in three consecutive protocol-controlled studies of Wilms' tumor at St Jude Children's Research Hospital. Tumors were completely excised at the time of diagnosis whenever possible, and patients were subsequently treated with chemotherapy and radiotherapy according to the initial extent of disease. All patients received dactinomycin and vincristine, with doxorubicin added to the regimens in studies 2 and 3. Chemotherapy was extended to 18 months in study 2 (n = 53), but was limited to 12 months for most patients in study 3 (n = 107). In the third study, radiation was eliminated altogether for patients with stage I or II tumors and was reduced to 12 Gy for those with more advanced disease. Intensification of chemotherapy in study 2 improved the 5-year relapse-free survival rate over that in study 1 (82% v 52%), but the accompanying increase in toxicity was considered unacceptable. Comparison of 2-year relapse-free survival rates in studies 2 and 3 indicated that the reduction of therapy in the latter trial did not jeopardize disease control: 88% v 86% for patients with stage II or III disease, favorable histology; 75% v 57% for the same stages, unfavorable histology; and 57% v 61% for stage IV patients. At least 80% of all patients enrolled in study 3 will be long-term survivors. We conclude that rescheduling of effective antitumor drugs and eliminating or reducing radiotherapy are feasible alternatives in the treatment of Wilms' tumor with favorable histologic features.