In the third National Wilms' Tumor Study (NWTS-3), patients with Stage II favorable histologic type (FH) or Stage III FH Wilms' tumor were randomized according to a factorial design for both radiation therapy (RT) and chemotherapy to be given after nephrectomy. Patients with Stage II FH disease were randomized between 2000 cGy and no postoperative RT; patients with Stage III FH disease were randomized between 2000 and 1000 cGy. No significant differences in survival were noticed. Although there were no significant differences in the rate of intraabdominal relapses, those patients with Stage III disease who received 1000 cGy and dactinomycin and vincristine (seven patients) experienced a relapse in the abdomen more frequently than those who received 2000 cGy and dactinomycin and vincristine (three patients), 1000 cGy and dactinomycin, vincristine, and doxorubicin (three patients), or 2000 cGy and dactinomycin, vincristine, and doxorubicin (two patients). This would suggest that doxorubicin might be a good substitute for the second 1000 cGy of RT. Boost doses of RT, although allowed, were rarely given and no assessment of the value of supplemental RT can be made. The dismal prognosis of abdominal relapse after RT is confirmed and delay of initiation of treatment beyond 10 days after surgery was a significant adverse factor as in NWTS-1 and NWTS-2.