Large-conduit coronary arteries respond to vasoactive stimuli differently than smaller coronary arterioies, but the quantitative effects of many vasoactive stimuli at various levels of the microvasculature remain unknown. To determine the site of constriction or dilation to serotonin and vasopressin in the coronary microcirculation, we studied microvascular responses in the left ventricle of anesthetized cats (n =36). To compensate for motion due to contraction of the heart, the epicardium was visualized with stroboscopic epi-illumination controlled by a computer to flash once per cardiac cycle in mid-diastole, making the vessels appear stationary. Serotonin (16 /ig/kg/min) or vasopressin (0.5 units/min) was infused into the left atrium while maintaining aortic pressure constant with a snare on the descending aorta or inferior vena cava. Myocardial blood flow was measured with radioactive microspheres. During infusion of serotonin, aortic pressure and heart rate did not change, but myocardial perfusion increased 90±38% (mean±SEM) from a control value of 159±27 ml/min • 100 g. Arteries and arterioies larger than 90 /im constricted in response to serotonin (control 159±12 fim; percent change -18±3; range -41 to 10%) while arterioies less than 90 /tin dilated to serotonin (control 54±7 fim; percent change 22 ±9; range -10 to 62%). During infusion of vasopressin, aortic pressure and heart rate did not change, and myocardial perfusion decreased 16±7% (control, 147±18 ml/min • 100 g). In contrast to serotonin, infusion of vasopressin constricted arterioies less than 90 /un (control, 55±5 fim; percent change -16±3; range -27 to -2%) while arteries and arterioies larger than 90 /im did not respond or dilated modestly (control, 190 ±11 fim; percent change 4±2; range -29 to 43%). These responses to serotonin and vasopressin suggest that vasomotor regulatory mechanisms vary in different size arteries and arterioies in the coronary microcirculation, and the pivotal size at which differential changes occur is the 90 -/im level. (Circulation Research 1989;65:343-351) P revious studies of the coronary microcirculation have been hampered by the inability to visualize arterioies because of respiratory and cardiac-induced motion of the heart. Techniques that have been used to study segmental coronary vascular resistance include measurements of pressure gradients and measurements of large