Ischemic heart disease is more apparent in the subendocardial than in subepicardial layers. We investigated coronary pressure-flow relations in layers of the isolated rat left ventricle, using 15 microm microspheres during diastolic and systolic arrest in the vasodilated coronary circulation. A special cannula allowed for selective determination of left main stem pressure-flow relations. Arterio-venous shunt flow was derived from microspheres in the venous effluent. We quantitatively investigated the pressure-flow relations in diastolic arrest (n=8), systolic arrest at normal contractility (n=8) and low contractility (n=6). In all three groups normal and large ventricular volume was studied. In diastolic arrest, at a perfusion pressure of 90 mmHg, subendocardial flow is larger than subepicardial flow, i.e., the endo/epi ratio is approximately 1.2. In systolic arrest the endo/epi ratio is approximately 0.3, and subendocardial flow and subepicardial flow are approximately 12% and approximately 55% of their values during diastolic arrest. The endo/epi ratio in diastolic arrest decreases with increasing perfusion pressure, while in systole the ratio increases. The slope of the pressure-flow relations, i.e., inverse of resistance, changes by a factor of approximately 5.3 in the subendocardium and by a factor approximately 2.2 in the subepicardium from diastole to systole. Lowering contractility affects subendocardial flow more than subepicardial flow, but both contractility and ventricular volume changes have only a limited effect on both subendocardial and subepicardial flow. The resistance (inverse of slope) of the total left main stem pressure-flow relation changes by a factor of approximately 3.4 from diastolic to systolic arrest. The zero-flow pressure increases from diastole to systole. Thus, coronary perfusion flow in diastolic arrest is larger than systolic arrest, with the largest difference in the subendocardium, as a result of layer dependent increases in vascular resistance and intercept pressure. Shunt flow is larger in diastolic than in systolic arrest, and increases with perfusion pressure. We conclude that changes in contractility and ventricular volume have a smaller effect on pressure-flow relations than diastolic-systolic differences. A synthesis of models accounting for the effect of cardiac contraction on perfusion is suggested.