In a control group of 32 patients undergoing open-heart operation, erythrocyte 2,3-diphosphoglycerate (2,3-DPG) declined progressively during the course of perfusion from a prebypass mean of 17.00 to 11.29 mu M per gram of hemoglobin at the end of bypass. The decrease was greater than that attributable merely to dilution of the patients' cells with the 2,3-DPG-deficient donor cells used to prime the pump oxygenator circuit. Administration of 300 mg of allopurinol, to prevent the conversion of inosine to uric acid, every 8 hours during the 24 hours prior to operation in 11 patients did not prevent the 2,3-DPG decrease during heart-lung bypass: prebypass, 18.31; postbypass, 13.56 mu M/gm Hgb. The mean P50 for both these groups combined decreased from a prebypass mean of 25.7 to a postbypass level of 21.9 torr. A solution of 0.1 M inosine, 0.1 M pyruvate, and 0.066 M phosphate (IPP) in a dosage of 7.5 ml per kologram of body weight prevented the 2,3-DPG decrease: prebypass, 15.74; postbypass, 14.85. Administration of 15 ml per kilogram of IPP in 15 patients preserved 2,3-DPG: prebypass, 18.09; postbypass, 18.52. The P50 remained unchanged in this last group. The method of providing for myocardial oxygen requirements during bypass was not standardized, and therefore the protective effect of IPP against ischemic damage in patients undergoing aortic valve replacement or myocardial revascularization could not be evaluated. No deleterious effects of IPP were noted.