A pericardial knock is a common finding in constrictive pericarditis. However, its origin has been uncertain. One hypothesis suggests that it is due to sudden deceleration of ventricular filling. To validate this hypothesis, left ventriculograms, phonocardiograms and external pulse recordings were obtained in seven patients with hemodynamic and pathologic findings of constrictive pericarditis and in seven normal subjects. Left ventriculographic silhouettes were digitized and left ventricular volumes were calculated by computer at 16 ms intervals. Curves of left ventricular volume versus diastolic filling time were constructed for each patient. Pericardial knock was recognized as an early high frequency sound recorded between 90 to 120 ms after the aortic closing sound and occurring at the trough of the Y descent of the jugular venous pressure tracing. The timing of the pericardial knock in five patients with constrictive pericarditis corresponded to a sudden and premature plateau of the diastolic left ventricular volume curve representing 85 +/- 4 percent (mean +/- standard deviation) of ventricular filling. The diastolic plateau was missing in two patients with constrictive pericarditis who had no pericardial knock. In these cases, the rate of ventricular filling was faster than normal in the first 20 percent of diastole. Thus, this study related pericardial knock to an abrupt plateau inthe diastolic left ventricular volume curve, supporting the view that sudden cessatin of ventricular filling generates the pericardial knock of constrictive pericarditis. Two mechanisms are proposed by which the filling plateau may produce the knock, and it is postulated that both ventricles may participate in the knock phenomenon.