The selection of the optimum treatment schedule has been one of the most vexing problems in high-pressure oxygen therapy. Because of the complexity of the procedures involved, it is desirable to administer therapy in one or very few sessions. On the other hand, a conventional treatment schedule of many fractions makes dose selection less critical and permits a more flexible approach that can be adapted to individual reactions. In addition, such a regimen would make it immediately possible to assess the advantages of high-pressure oxygen against extensive experiences gained in conventional tumor therapy. The overriding question, however, is whether the oxygen-mediated enhanced sensitivity of malignant cells is more effectively exploited by one or by several treatments if the assumption is made that the sensitivity of normal cells is unchanged or, at least, less changed by increased oxygen tension. This central question appears to be largely unresolved, and it is therefore not surprising that investigators employing hyperbaric oxygen therapy use a variety of schedules and commonly alter them. An extensive literature deals with the theoretical aspects of fractionation in radiotherapy. Much of this is cited in a recent paper by Wootton (1) in which the subject is also considered in relation to oxygen therapy. Most of the modern theoretical considerations are developments of an approach originated by Elkind (2). Using a simple mathematical model and applying the findings of his extensive research into recovery mechanisms, Elkind proceeded to derive a variety of relations between treatment efficacy and fractionation for various parameters of the survival curve. The latter was assumed to be expressed by This formula represents a model in which cellular inactivation occurs after m independent sites within the cell have been inactivated, the inactivation rate of each site being 1/D0- It is generally agreed that Equation (1) is only an approximation to the actual shape of the dose-effect curve, and indeed there are reasons which indicate that it is incorrect, both at low and at high doses (3–5). As pointed out by Elkind, however, minor deviations are of little significance. In his work as well as in this communication the intent is not to provide accurate numerical answers, but merely to indicate general trends that are significant. Following Elkind, we will make the assumption that if radiation is administered at conventional intervals (i.e., daily or less frequent), sublethally injured cells recover completely in the interim and the surviving population follows the same inactivation curve as the original one. Thus if the total dose Df is administered in f fractions, the survival, Cf, is given by