Quantitative data obtained from different groups of subjects are in agreement and make it possible to evaluate the risk of leukemia following total body irradiation with doses exceeding 100 rads of X-rays or lower doses of neutrons. Irradiations which are more localized or administered over a longer period of time retain their leukemogenic effect when the total dose exceeds a few hundred rads. The dose-effect relationship appears to be curvilinear and, from the scientific standpoint, linear extrapolation to predict a low dose effect is not justified. Such linear extrapolation is nevertheless legitimate on grounds of radioprotection but should not be used to evaluate the risk of small doses such as those given for radiodiagnostic purposes (doses in the range of one rad), or those received by a population in the vicinity of a nuclear power station (doses in the range of one rad), or those received by a population in the vicinity of a nuclear power station (doses in the range of a millirad). Combined radio-chemotherapy modalities, as used in cancer treatment and particularly for malignant lymphoma, carry a high risk of inducing secondary leukemias. These combined treatment regimens should be limited to cases for which radiotherapy alone would be insufficient.