Nuclear deoxyribonucleic acid ploidy studies with paraffin-embedded archival tumor specimens were performed by flow cytometry on extracted nuclei from 56 Wilms tumors. Before the era of chemotherapy 9 patients had a 33 per cent survival rate at 5 years. No significant correlation between deoxyribonucleic acid ploidy pattern and survival was seen in this early group of patients. Since 1960, 47 patients underwent radical nephrectomy and received chemotherapy. Deoxyribonucleic acid histograms in this group were normal in 13 tumors, while 23 tumors exhibited a tetraploid pattern (greater than 10 per cent nuclei in the G2 peak) and 11 showed an aneuploid pattern. Survival rates according to deoxyribonucleic acid ploidy pattern showed that 100 per cent of the patients with a normal or aneuploid histogram pattern were alive at 2 and 5 years. Patients with a deoxyribonucleic acid tetraploid pattern had a significantly worse survival of 74 and 69 per cent at 2 and 5 years (p less than 0.02 and less than 0.01), respectively. Tumors with a normal or aneuploid histogram pattern for all stages and a deoxyribonucleic acid tetraploid pattern for pathological stage I or II were associated with a 2 and 5-year survival rate of 100 and 97 per cent, respectively. However, patients with stages III and IV tumors with a deoxyribonucleic acid tetraploid pattern had a significantly worse survival of 25 per cent at 2 and 5 years (p less than 0.0001). Measurement of nuclear deoxyribonucleic acid ploidy identifies a patient group who are at high risk for treatment failure, that is stages III and IV tumors with a deoxyribonucleic acid tetraploid pattern.