The effects of bronchial arterial administration of vasoactive substances on the pulmonary circulation were studied by a new technique for selective catheterization of a bronchial artery in intact dogs. In most experiments, this technique permitted pressor agents to be distributed mainly to one lung with smaller amounts to the other lung. The intercostal arteries were avoided, and in all but 2 of 23 experiments only microscopic quantities of injected India ink could be identified in the distribution of the esophageal and mediastinal branches. These studies indicate that serotonin, angiotensin, histamine, and norepinephrine injected selectively into a bronchial artery increase lobar arterial pressure. Since blood flow was constant and left atrial pressure did not change, the increase in pressure suggests active pulmonary vasoconstriction. Additionally, the responses to bronchial and lobar arterial injections of pressor agents were similar. The contribution of bronchopulmonary shunt flow to pulmonary flow was small, since, under conditions of controlled lobar blood flow, changes in bronchial flow elicited by 65-75-mm Hg changes in bronchial arterial pressure produced little if any change in pressure in the perfused lobar artery or small vein. Bronchoconstriction contributed little to the response to bronchial administration of pressor agents, since responses were similar in the ventilated and the collapsed lobe. Injection of vasoflavine dyes into the bronchial artery showed the close proximity of bronchial and pulmonary arteries and confirmed the bronchial arterial origin of the vasa vasorum of pulmonary arteries. No vasa venorum were identified. Although no direct anatomic bronchial artery-pulmonary artery shunt was identified, ascorbic acid and 5-hydroxydopamine diffused rapidly into intrapulmonary arteries from the bronchial artery. These data suggest that the pulmonary pressor response results from passage of the vasoactive agents from the bronchial artery to the lobar artery through the vasa vasorum and by diffusion. Since no vasa venorum were found, pulmonary venoconstriction probably resulted from pressor agents reaching the veins by way of bronchopulmonary shunt flow. These results suggest a mechanism by which pressor substances present or liberated in the bronchial vascular bed can affect tone in the pulmonary vascular bed.