The relationship between blood flow and differential pressure in the main pulmonary artery, expressed as longitudinal impedance, was determined in anaesthetized open-chest dogs and compared with the impedances predicted by the Womersley theory. Blood flow was measured with an electromagnetic flowmeter, and differential pressure with double-lumen catheter or by needles piercing the vessel wall. For the first five harmonics (α = 10 to 35), the measured impedance modulus was greater than the theoretical value by an average of 18%, and flow lagged differential pressure by an average of 0·53 radians more than theory predicted. The ratio of experimental to Womersley values averaged 9.6 for the resistive component of impedance, and 0·89 for the reactive component. These discrepancies between theory and observation, which presumably arise because conditions of flow in the pulmonary artery differ from those assumed in Womersley's analysis, were qualitatively consistent with the effects expected from tapering of the artery, non-laminar blood flow, and wall viscosity.