SUMMARY We divised a method to determine tissue osmolality in intact beating hearts. After occlusion of the left anterior descending coronary artery (LAD) of isolated porcine hearts, tissue osmolality in the ischemic myocardium increased within 50 minutes by about 40 mOsm/kg. This rise in osmolality could be accounted for by metabolic processes, notably the conversion of glycogen into lactate, and the hydrolysis of high energy phosphates. Concomitant with the rise in osmolality, the ischemic myocardium during the 1-hour period of LAD occlusion took up fluid and increased tistue water volume by an average of 16.5%. We demonstrated that the osmolality of fixatives used for morphological studies markedly influences ischemic cell morphology. Thus, normotonic fixation of the ischemic myocardium accentuates cell swelling, whereas nearly normal cell volumes result from hypertonic fixation, adjusted according to the rise hi ischemic tissue osmolality. Normotonic reperfusion of the ischemic area after 1 hour of LAD occlusion resulted in the "no-refiow" phenomenon in the mldmural and subendocardial regions. Epicardial and intramural DC-electrograms showed persistent ischemic changes, i.e., T-Q depression, S-T elevation, and monophasic potentials. Tissue resistivity, which during ischemia had risen twofold, remained high. Lacate levels remained high, creatinephosphate (CP) and adenosinetriphosphate (ATP) levels remained low. Selective hypertonic reperfusion of the LAD, followed by a gradual return to normotonic perfnsion, resulted in a normalization of DC extracellular elcctrograms, restoration of electrical resistivity to near normal, low levels of lactate, and higher levels of CP and ATP although control values were not reached. Cell morphology was correspondingly normalized following this procedure. We conclude that ischemic cells become hyperosmotic and consequently take up additional fluid when exposed to normotonic blood. This increased cell swelling compresses capillaries, prevents reperfusion, and may be a major factor in causing reperfusion damage. This damage can be prevented to a large extent by selective hypertonic reperfusion. Circ Res 49: 364-381, 1881