Assessment of coronary blood flow and the vasodilator reserve capacity of individual coronary arteries in the catheterization laboratory has been hampered by methodologic limitations. We have developed and validated a small Doppler catheter that can subselectively measure phasic coronary blood flow velocity (CBFV). In seven anesthetized calves, CBFV was varied from 0.1 to 5.7 times control CBFV. Changes in mean CBFV measured intraluminally by catheter in the left anterior descending and left circumflex arteries were similar to those measured simultaneously with an epicardial Doppler probe on the surface of the same vessel (n = 85, r = .95, slope = 1.04) and to changes in coronary sinus flow (n = 69, r = .97, slope = 1.06) measured with timed venous collections. Identical maximal coronary reactive hyperemic responses with the catheter present and absent in the artery being studied demonstrated that coronary obstruction by the catheter was minimal. Safety studies in six additional calves demonstrated that the catheter caused small changes in coronary endothelial permeability. Histologic studies revealed no endothelial denudation or thrombus formation. Stable phasic recordings of coronary blood flow velocity have been obtained in 58 of 70 patients studied. One of the 70 patients studied had abrupt coronary occlusion probably related to catheter-induced vasospasm. In 10 normal patients, intracoronary meglumine diatrizoate increased CBFV to 3.5 times that at rest (range 2.8 to 5.0). CBFV rose 5.0-fold after an intravenous infusion of dipyridamole (range 3.8 to 7.0). In each patient, dipyridamole produced greater vasodilation than meglumine diatrizoate. The time- and dose-response characteristics to dipyridamole infusion were heterogeneous, underscoring the advantage of continuous on-line measurement of CBFV in the measurement of vasodilator reserve. This method of measuring CBFV and assessing vasodilator reserve in the catheterization laboratory should facilitate studies of the coronary circulation in man.