T HE OXYGEN consumption of the heart has been variously correlated with the product of heart rate and mean aortic pressure,’ the “tension-time” index,2 and the developed tension3 defined as the product of pressure, cube root of the output (radius) and heart rate. The basal requirement of the fibrillating heart, which like the living asystolic heart is performing no external work, has also been reported as having variously a higher4-7 or the same* oxygen consumption as the empty beating heart and the potassiumor vagal-arrested heart. There is also a wide belief among cardiac surgeons that the fibrillating heart has a lower oxygen requirement than the nonworking beating heart.9 In addition to potassium arrest, it is possible to produce mechanical asystole while keeping unchanged the function of initiating and propagating impulses, the membrane effect producing the electrocardiogram.1° This mechanical asystole can be produced by perfusing the preparation with a calcium-free solution. When magnesium is substituted for calcium in the perfusion fluid, electrocardiographic manifestations of calcium deficiency are prevented, although this ion produces no effect on the return of contraction. Magnesium is capable of substituting for calcium in membrane activity but not in excitation-contraction coupling. Oxygen consumption of the isolated heart has been measured under these conditions of electromechanical dissociation. METHOD Isolated rabbit hearts were perfused on an Anderson apparatus.” After ether anesthesia the chest was opened, the aorta cannulated and perfusion initiated before the heart was removed. Both superior venae cavae and the inferior vena cava were then ligated, and the isolated heart was transferred to the perfusion system. Electrocardiograms were recorded by means of hook electrodes and a Sanborn Polyviso. An electronic pacemaker attached to the A-V node drove the heart at a selected rate of 120/min., or as desired. Ventricular fibrillation was induced electrically. The perfusion fluid was Chenoweth’s solution,‘* through which was constantly bubbled a mixture of 95 per cent 02 and 5 per cent COZ. The experimental pet&sate was identical except that mag nesium replaced the calcium on a molar basis. All perfusion solutions were kept at a temperature of 39’ c., and the entire system was surrounded by a water jacket at the same temperature. The mixed coronary sample was taken from a needle in the right ventricle. Since the systemic veins entering the right atrium were tied off, fluid entering the right ventricle came only from the coronary sinus and coronary veins and was not in contact with air at any time. The arterial sample was obtained from the aortic cannula. The oxygen content of these samples was determined on a Van Slyke gasometric apparatus. The arterial oxygen content varied from 1.75 to 2.1 vol.‘%, and the venous oxygen from 0.5 to 2.0 vol.%. The sample size averaged 5 cc., except in cases of low coronary flow where 3 cc. was used. All duplicate determinations checked to within 0.05 vol.oje. The coronary venous return was pumped by the right