In order to evaluate the functional characteristics of the intramyocardial capacitance vessels during prolonged diastole, we analyzed the response of coronary vein flow after stepwise changes of coronary artery pressure in anesthetized open-chest dogs by using our newly developed laser Doppler velocimeter with an optical fiber. The peripheral portion of the great cardiac vein was isolated and the optical fiber tip was inserted into the vessel. The left anterior descending coronary artery was cannulated and connected to a reservoir to regulate coronary perfusion pressure. Intracoronary adenosine administration was carried out to avoid any change in coronary vasomotor tone. After 15 seconds of occlusion of the perfusion route, the heart was arrested by pacing-off. Two seconds later, coronary perfusion pressure was increased stepwise to a preset target pressure. This procedure was repeated by changing target pressure at 4 (or 5) different pressure levels (31-105 mm Hg). The great cardiac vein flow became zero due to the cardiac arrest and remained at zero for a moment (dead time) after the initiation of reperfusion. Then the flow reappeared and increased with first order time delay. The presence of dead time indicates the existence of unstressed volume, and the first order time delay represents the product of resistance and capacitance. The unstressed volume with a minimal vasomotor tone for perfusion pressure of 60-90 mm Hg was 5.2 +/- 2.2 ml per 100 g left ventricle, which is comparable to coronary blood flow for several beats. The capacitance at perfusion pressure of 60-90 mm Hg was 0.08 +/- 0.04 ml/mm Hg per 100 g left ventricle, while that at low perfusion pressure (30-50 mm Hg) was 0.14 +/- 0.09 ml/mm Hg per 100 g left ventricle. These results indicate that the intramyocardial capacitance vessels have two functional components, and that the phasic nature of coronary vein flow is solely the result of the myocardial squeezing of the blood in the capacitance vessels.