Results from several laboratories, including ours, have suggested that measurements of radiation-induced DNA strand breaks and DNA-protein crosslinks (DPCs) may be used to estimate the hypoxic fraction or fractional hypoxic volume of tumors and normal tissues. This suggestion has been predicted on both published and unpublished information that (1) the oxygen dependence of the formation of strand breaks in irradiated mammalian cells is similar to the oxygen dependence of radiation-induced cell killing, and (2) the oxygen dependence of the formation of DPCs in irradiated mammalian cells is the mirror image of the oxygen dependence of radiation-induced cell killing. However, the published studies that attempted to determine the relationship between the oxygen dependence of the formation of strand breaks and the radiation sensitivity of mammalian cells were not performed at 37 degrees C, the exact oxygen concentrations were not always known, and the results were conflicting. In addition, most of the data on the oxygen dependence of the formation of DPCs are unpublished. Consequently, we have undertaken a comprehensive investigation of one cell line, 9L/Ro rat brain tumor cells, to determine if the shape of the oxygen dependence curve and the Km value for radiation-induced strand breaks and DPCs were similar when 9L cells were irradiated under both ideal gas-liquid equilibrium conditions at 4 degrees C and nonideal gas-liquid equilibrium conditions at 37 degrees C. At 4 degrees C under ideal gas-liquid equilibrium conditions, the Km for the formation of strand breaks was approximately 0.0045 mM, and the Km for radiation sensitivity was approximately 0.005 mM. A similar comparison for the formation of DPCs at 4 degrees C could not be made, because the efficiency of the formation of DPCs was much lower at 4 degrees C than at 37 degrees C. At 37 degrees C under nonideal gas-liquid equilibrium conditions, the apparent Km for the formation of strand breaks and radiation sensitivity was approximately 0.032 mM, and the Km for the formation of DPCs was approximately 0.02 mM. The data for strand breaks are in agreement with the published data of Chapman et al. (Int. J. Radiat. Biol. 26, 383-389, 1974), and the data for DPCs are in agreement with the unpublished data of Meyn (personal communication). These results support the suggestion that measurements of radiation-induced strand breaks and/or DPCs may be used to detect hypoxic cells and estimate the hypoxic fraction or fractional hypoxic volume of tumors and normal tissues.