We tested the hypothesis that reductions of perfusion pressure distal to a flow-limiting coronary artery stenosis can directly impair perfusion of the subendocardial myocardium. Dogs were instrumented with an electromagnetic flowmeter probe and a variable occluder on the proximal left circumflex coronary artery. Coronary perfusion pressure was measured with a catheter distal to the occluder. Coronary autoregulation was abolished by intraarterial infusion of adenosine to produce maximal coronary vasodilation. The transmural distribution of myocardial blood flow was measured with radioactive microspheres during unimpeded arterial inflow, when the occluder was progressively narrowed to reduce distal coronary pressure to approximately 70%, 50% and 35% of the control coronary perfusion pressure, and during total coronary occlusion. Heart rate, left ventricular diastolic pressure and the fraction of coronary artery flow during systole remained constant throughout the study. Progressive reductions of coronary perfusion pressure were accompanied by direct reductions of the subendocardial/subepicardial blood flow ratio (r=0.83). Examination of the relationship between myocardial blood flow and coronary perfusion pressure showed that blood flow decreased linearly with perfusion pressure, with flow ceasing at a positive pressure (zero-flow pressure). Blood flow data from four transmural myocardial layers from epicardium to endocardium showed that this zero-flow pressure increased progressively from 10 ± 2.1 mm Hg in the subepicardium to 18 ± 2.3 mm Hg in the subendocardium (p<0.01). Consequently, as coronary pressure was reduced, the zero-flow pressure represented a progressively greater fraction of coronary pressure in the subendocardium than in the subepicardium. This effect appeared to account for the progressive redistribution of blood flow away from the subendocardium that occurred as coronary pressure was decreased. Myocardial vascular resistance did not change as a result of changes in coronary perfusion pressure.