The design of the study was to determine whether an increased blood flow as seen in shunt lesions could serve as a stimulus for the secretion of atrial natriuretic factor (ANF). Since atrial pressure, flow, and dilatation are closely related, an experimental ductus arteriosus model was utilized, in which acute changes of flow are assumed not to dilate the left atrium. In six dogs, a Dacron graft was constructed between the main pulmonary artery and the innominate artery. Constricting and releasing the tape around the graft adjusted the amount of "ductal" shunting. The total pulmonary flow and the shunt flow were measured by electromagnetic-flow transducers around the aortic root and around the graft. Plasma ANF concentration was measured from both cardiac atria. The size of the left atrium was determined from echocardiographic measurements made from a short-axis view. The total pulmonary flow varied between 1.2 and 5.8 1/min. The highest measured ANF was 396 pg/ml, and this was from the left atrium when the pressure was 18 mmHg, the highest left atrial pressure recorded. The highest right atrial pressure (5 mmHg) also correlated with the highest right-atrial level of ANF (366 pg/ml). The right atrial pressure had a significant correlation with plasma ANF concentration (R = 0.43, p less than 0.05). Pulmonary flow and plasma ANF concentration did not correlate; neither did left atrial size and ANF levels in 16 flow states where the size was measured. In the absence of atrial dilatation there was minimal stimulus for ANF secretion. A transient increase of left atrial pressure, without a concomitant significant atrial dilatation, did not serve as a significant stimulus for ANF secretion.