In 21 dogs on cardiopulmonary bypass with ventricles kept empty, the mean beating but nonworking myocardial oxygen consumption (mVO2) was 3.8 ml/min per 100 g at a heart rate of 158 beats/min. After subsequent potassium chloride arrest, the basal mVOs was 1.74 ml/min per 100 g. To compare these values with the working situation, we measured mVOj in these hearts before instituting bypass when the heart rate averaged 179 beats/min and arterial pressure averaged 108 mm Hg; mVOj was 9.2 ml/min per 100 g. Atrioventricular dissociation was induced in five beating, nonworking hearts; electrical pacing at increasing heart rates produced a linear increase in mVO2. Extrapolation to zero heart rate yielded a value of 1.24 ml/min per 100 g, which was not significantly different from the KC1 arrest value of 1.25 ml/min per 100 g in these same hearts. These measurements permitted calculation of an energy expenditure per beat of 10.4 mj/g. Subtraction of the basal value of 2.0 mJ/g yielded a value for energy of contraction of 8.4 mJ/g. The stress-independent component was estimated to be 2.7 mJ/g. The basal rnVO2at normal perfusion pressure remained constant for arrest periods of at least 1 hour. Perfusion with modified Krebs-Henseleit buffer caused a 30% reduction in the basal mVOj in spite of maintained perfusion pressure, O2 supply in excess of consumption, coronary venous Po2values above 60 mm Hg, and the addition of amino acids. It is possible that this finding indi-cates deficient oxygenation of myocardial cells in hearts perfused with solutions lacking red cells. Circ Res 47: 408-417, 1980