The concomitant changes in components of external cardiac power and phasic coronary flow are described here in an aorto-caval fistula model (ACF) in the dog. Eleven animals were used in this study. The steady, oscillatory, and kinetic components of the external cardiac power and aortic impedance spectrum were calculated from the measurement of instantaneous flow and pressure in the ascending aorta. Arterial coronary flow was measured by a pulsed doppler flowmeter on the left anterior descending artery. The concomitant venous coronary flow was measured in the last two dogs. After unclamping of a large aorto caval fistula, mean blood pressure decreased from 96 ± 31 mm Hg in the control state to 78 ± 19 mm Hg 5 min after opening the shunt. Aortic blood flow and heart rate increased significantly. These changes in pressure and flow led to a decrease in the steady component of the external cardiac power (752 ± 203 mW in the control state to 590 ± 180 mW in the ACF; P < 0.01), whereas the oscillatory and the kinetic components of the total external power increased (23.5 ± 11 mW in the control state to 33 ± 6 mW in the ACF for oscillatory power; 11.7 ± 5.1 mW in the control state to 17.7 ± 8.5 mW in the ACF for kinetic power, P < 0.05). This increase in the oscillatory component of the cardiac power corresponded to the increase in the characteristic impedance of the aorta in the ACF (2,507 ± 231 dyn · sec · cm −5 in the control state versus 4,366 ± 745 dyn · sec · cm −5 in the ACF). These results suggested poor energetic efficiency in the matching between the left ventricle and the large arteries in this model. After unclamping the fistula, the systolic component of arterial coronary inflow becomes biphasic with a constant reverse flow from the myocardium to the aorta. The diastolic part of coronary arterial inflow decreased markedly, secondary to the decrease in perfusion pressure, and then reincreased to values near control values (59 ± 22 ml/min in the control state versus 52 ± 18 ml/min 5 min after unclamping the shunt). In five cases, the abrupt fall in blood pressure induced by unclamping the fistula led to a complete disappearance of diastolic coronary arterial inflow in diastoles of normal duration. The measured pressure of this zero inflow was about 30 mm Hg. During this zero coronary inflow, venous coronary outflow decreased but never disappeared. Several diastoles later, while perfusion pressure remained very low, coronary inflow and outflow increased consequent to autoregulatory vasodilatation. These results obtained in diastoles of normal duration demonstrated a significant role of intramyocardial capacitance in the definition of coronary autoregulation.