The accumulating evidence from our studies of the plasma volume in hemorrhagic shock agrees with that of Gregersen (1) and of Evans (2), to the effect that there is no further loss of plasma from the circulation after the initial loss caused by the bleeding. Radioactive plasma proteins in the tissues of the dog in hemorrhagic shock are not found in greater concentration than in the unshocked dog. The irreversibility of hemorrhagic shock therefore cannot be explained by the theory of increased capillary permeability.Venous anoxemia is one of the outspoken pathological phenomena in shock. It appears early, and the associated tissue anoxia, if prolonged sufficiently, might be considered responsible for the development of irreversibility because of irretrievable damage to vital structures. In that case, the prevention or amelioration of venous anoxemia might exert a beneficial therapeutic effect. Indeed, several recent reports (3,4) utilized with the expectation of better saturation, primarily because of increased oxygen in physical solution in the circulating plasma (5). Because four atmospheres of oxygen is lethal for dogs if carried on for the time intervals required for shock experiments, we performed all our experiments at three atmospheres of oxygen, in a pressure chamber made available to us by Dr. Drinker at the Harvard School of Public Health. Temperature and humidity were controlled throughout the experiment in the chamber and decompression carried out so as to avoid any ill effects from "the bends." Two control experiments were performed to observe possible deleterious effects of breathing three atmospheres of oxygen for a period of three hours-the maximum interval utilized in this study-in dogs prepared according to the technique described below, except that no blood was withdrawn. The dogs showed no effect on the blood pressure or in any other respect and recovery without incident ensued.Mongrel dogs were anesthetized according to Wiggers' technique (6), i.e., about one-half hour following an initial dose of morphine sulphate (3.0 mgm. per kgm.) subcutaneously, sodium barbital (175 mgm. per kgm.) was given intravenously, two thirds of the dose in one injection and only as much more of the remaining third as was necessary for minimum anesthesia, sufficient to permit tracheal cannulation and exposure of vessels. No further anesthetic was required, although the lid reflex was present and many dogs were on the verge of recovering consciousness in the later stages. Blood was withdrawn fractionally from the femoral artery over a period of one-half to 1% hours, until the arterial blood pressure fell to 70 mm. Hg and did not rise within fifteen minutes after the last withdrawal. No further blood, except for sampling, was removed, regardless of any subsequent recovery of blood pressure. In all, some 30 to 40 per cent of the estimated total blood volume was withdrawn. Plasma volumes before and after bleeding were measured by the method of Gibson and Evelyn (7).Hematocrits were determined in Wintrobe tubes contai...