For several years, we have concentrated our efforts on validating the use of radiation-induced DNA strand breaks and DNA-protein crosslinks to assess the oxygenation status of tumors and normal tissues. We have demonstrated that (1) the oxygen dependence of strand break formation is identical to that of radiation-induced cell killing; (2) the oxygen dependence of DNA-protein crosslink formation is the mirror image of that of radiation-induced cell killing; and (3) the formation of these radiation-induced DNA lesions is predominantly dependent on the oxygen concentration near the DNA and is independent of the cell type, metabolic status, proliferative status, pH of the surrounding environment, and composition or properties of the proteins tightly associated with the DNA. In the present study, the hypoxic fraction of three experimental tumors was estimated using our assay of radiation-induced DNA damage. The average hypoxic fraction of a large number of tumors estimated with this assay of radiation-induced DNA damage for (1) WiDR human colorectal carcinoma xenografts (40.8 +/- 4.2%), (2) 66 mouse mammary adenocarcinoma tumors (41.8 +/- 3.1%), and (3) subcutaneous tumors grown from 9L rat brain tumor cells (95% CI =-8.2-4.2%) was not statistically different from that of a large number of tumors measured for each of these tumor models by the paired survival curve method (38.3 +/- 6. 3%, 28.9 +/- 5.5%, 95% CI = 2.2-4.4%, respectively). When the hypoxic fraction measured by the alkaline elution method on one half of an individual tumor was compared to that measured by the paired survival curve method on the other half of the same tumor, no statistical correlation was found for either 66 or WiDR tumors. Although this assay of radiation-induced DNA damage can be used effectively in the laboratory to answer a number of important questions about the oxygenation status of animal tumors and normal tissues, failure to reliably estimate the hypoxic fraction of individual tumors and technical considerations make it unlikely that the assay can be used in the clinic to estimate the hypoxic fraction of human tumors.