An animal model for the induction of coronary artery stenosis is described. In this model the degree of stenosis, as induced with commercially available hydraulic occluders, can be easily controlled by keeping constant the mean perfusion pressure (pcor) distal to the site of stenosis. This pcor is the input signal for a servo-system feeding a motor-pump, which determines the degree of inflation of the cuff around the left anterior interventricular coronary artery (LAICA). In each experiment pcor did not vary more than 2 mm Hg from the preset value of about 25 mm Hg. In 60 anesthetized open-chest dogs the time course of standard hemodynamic variables and regional myocardial blood flow in the center of the underperfused area, using the radioactive microsphere technique, were determined. Within 1 min after induction of stenosis heart rate and end-diastolic left ventricular pressure (plved) increased (by 20 and 60%, respectively) and mean aortic pressure and dplv/dtmax decreased (by 10 and 25%, respectively). After the initial decrease median myocardial blood flow further decreased between 1 and 5 min of stenosis from 0.63 to 0.32 ml.min-1.g-1 in the outer layers (P less than 0.05) and from 0.26 to 0.15 ml.min-1.g-1 in the inner layers (P less than 0.05), despite constant hemodynamic conditions and pcor. Between 5 and 120 min of stenosis these values remained unchanged in the outer layers, but decreased further in the inner layers to 0.08 ml.min-1.g-1 (P less than 0.05). The accurate control of pcor, the reproducibility of the levels of residual blood flow and the ease of handling the stenosis system indicate that coronary artery stenosis controlled by perfusion pressure distal to the stenosis is a useful animal model to study events during regional myocardial ischemia. With the use of this model of low flow ischemia a biphasic increase of myocardial vascular resistance was observed, which is initiated during the first minutes of coronary artery stenosis.