In the double-strand break (DSB) model, the initial slope (alpha) of the survival curve is determined by the number of DSBs remaining unrepaired at a certain time after irradiation. There are three types of calculations which confirm the validity of this approach. First, the model quite reasonably predicts the low-LET alpha values for Chinese hamster V79, mouse L5178Y-R, normal fibroblast, ataxia telangiectasia (AT), and Ehrlich ascites tumor (EAT) cells using the measured values of the cellular DNA content, the radiation yield for DSB induction, the time constant for DSB repair, and the time available for DSB repair. Second, this approach correctly predicts an increase in alpha values for V79 and Chinese hamster ovary (CHO) cells due to postirradiation treatments with hypertonic saline or a high concentration of araA. These calculations are based on the duration of the DSB repair block induced by those treatments. Third, a fair agreement between the predicted and measured initial slopes for 3- to 4-MeV alpha particles was obtained for normal fibroblast, EAT, V79, CHO, Chinese hamster HS-23, C3H 10T1/2, and lung epithelial cells. However, the alpha values for AT and xrs-6 cells were overestimated. The calculation of high-LET alpha values uses the experimental values of the relative increase in the DSB repair time constant and in the radiation yield for DSB induction for alpha particles in comparison with those for X rays.