The effect of oxygen concentration on the radiosensitivity of cells has been extensively investigated in bacteria, plants, and lower animals (1,2). Nearly all the biological materials studied have shown that, as oxygen concentration is increased from anoxia to that of air, there is a sharp, approximately threefold, increase in radiosensitivity; but sensitivity is little enhanced, if at all, with higher concentrations. This “oxygen effect” appeared, in theory, to offer some usefulness for radiotherapy. The rationale has been presented by L. H. Gray and his associates in the British Journal of Radiology (1). In brief, it is as follows: Tumors are generally more anoxic than the normal tissues around them. Thus a tumor, since it is anoxic, occupies a low position on the oxy-gen-radiosensitivity curve as compared with normal, well vascularized tissue, which stands high on the radiosensitivity curve and, most important, is close to the plateau of maximum sensitivity. It might be possible, therefore, to increase the tumor radiosensitivity relative to that of normal tissue, by subjecting the whole animal to increased oxygen tension. On the basis of this possibility, it was important to establish that the radiosensitivity of mammalian cells was, in fact, affected qualitatively and quantitatively in the same way as the organisms previously investigated. The experimentally favorable material, Ehrlich mouse ascites tumor, was used to obtain data for mammalian cells. It grows rapidly as a free-cell suspension in the peritoneal cavity, is easily transferred, and gives 100 per cent takes. The strain used in one study was kindly furnished by Dr. George Klein (3). Ascites fluid containing the rapidly dividing tumor cells was withdrawn from animals inoculated six days previously, and was diluted with about one-fourth its volume of dextrose-saline plus heparin to prevent clotting. The fluid was equilibrated with atmospheres of various oxygen concentration by passing a jet of the gas, which stirred the fluid, over the surface for ten minutes before and during the in vitro irradiation. Since the tumor cells were freely suspended in the fluid, we can be quite certain that the cells themselves were in equilibrium with the gas. After irradiation, the fluid was re-inoculated into fresh mice, and then examined periodically in samples drawn by peritoneal puncture. In the cytological experiments, which will be described first, small samples of the irradiated fluid were withdrawn at various intervals following inoculation. The cells were fixed and stained, and the proportion of dividing cells showing abnormal anaphase division, that is, cells in anaphase which had chromosomal bridges, chromosomal fragments, or both, was determined. The prediction would be made, on perfectly sound genetic grounds, that such abnormal anaphase divisions would produce inviable daughter cells.