We examined pulmonary venous flow (PVF) velocity, pulmonary venous dimension (PVD) and transmitral flow (TMF) velocity during changing loading conditions in 15 anesthetized closed-chest dogs by transesophageal Doppler echocardiography (TEE). We also simultaneously measured PVF velocity and pulmonary venous volume (PVV) in opened-chest dog. TMF and PVF were satisfactorily obtained during changing loading conditions in all experimental dogs, and PVD could be accurately measured by TEE. TMF consisted of peak rapid filling flow velocity (R wave) and peak atrial flow velocity (A wave). PVF consisted of systolic forward flow (S wave) and diastolic forward flow (D wave). The following variables were measured in TMF and PVF: peak flow velocity of each wave (R, A, S, D), time-velocity integral of each wave (RI, AI, SI, DI). There was a significant correlation between the changes in RI and SI/DI during preload reduction (r=0.82, p<0.001) or afterload increase (r=-0.59, p<0.05). PVV was decreased by more than 50%, nevertheless PVF velocity was unchanged. Accordingly the discrepancy was admitted between volume and velocity of pulmonary vein during preload reduction. These results suggested that the changes in RI during changes in loading conditions might be attributed to the changes in left atrial reservoir volume and conduit volume. For the assesment of pulmonary venous return at different loading conditions, it is not enough to evaluate PVF velocity alone, and it should be paied attention to the changes in PVD.