During recent years the use of plasma transfusions has become widespread and increasingly so of late as a result of war conditions. More and more the need for really large amounts of plasma in treating certain types of cases is being appreciated. The data presented in this paper should be of interest to those engaged in studying the quantitative aspects of such therapy. Further these data have certain theoretical impfications that bear upon the more general problems of plasma protein formation and protein metabolism.Hyperproteinemia can regularly be produced in dogs, and presumably in other species of animals, by repeated injections of plasma obtained from homologous donors. Almost none of the injected protein escapes in the urine as protein, nor is it eliminated quantitatively as increased urinary or fecal N2. While no direct measurements have been made in dogs, Addis' results following intraperitoneal injections of serum in rats (1) indicate that the protein content of all of the viscera and tissues is increased. The greatest increment is in the serum, next in the liver, but all organs and tissues show a definite increase in protein content.Practically all of the experiments to date (2-6) have been carried out with the recipient of the plasma or serum injections maintained in a fasting state or receiving only sugar, or sugar and fat, by mouth. A priori one would expect that plasma injections in an animal maintained on a full diet or a high protein diet would yield summation effects and result in a more marked hyperproteinemia. In the experiments reported below this does not happen; in fact, a significantly higher hyperproteinemia is produced when the animals are maintained on a low protein diet than when a high protein diet is given. This seeming paradox is open to a number of interpretations but all of them seem to imply an equilibrium between food, plasma, and tissue proteins.