In an attempt to predict peak pulmonary artery pressure from routine M-mode echocardiographic tracings, 95 infants and children with congenital heart disease were examined. Following the Burstin method for prediction of peak pulmonary artery pressure, which was originally based on the phonocardiogram and jugular phlebogram, M-mode echocardiography was used to measure the interval from pulmonary valve closure to tricuspid valve opening, namely, the period of isovolumic diastole. The measured interval was plotted on a modified table relating the interval, heart rate and predicted peak pulmonary artery pressure. The peak pulmonary artery pressure predicted by echocardiography was compared with that measured at cardiac catheterization. The correlation between predicted and actual peak pulmonary artery pressure was good (r = 0.86) for routine studies with the patient in the nonsedated state. All patients with a predicted peak pressure less than 40 mm Hg were found at catheterization to have a pressure less than 40 mm Hg. The correlation was better (r = 0.96) when comparing predictions made from the echocardiogram obtained while the patient was sedated for catheterization. Prediction of the magnitude of elevation of peak pressure was especially good when prediction and measurement were nearly simultaneous. Predictions were less accurate in the presence of tachycardia at rates of more than 155 beats/min. The method for estimating peak pulmonary artery pressure from M-mode echocardiographic tracings is reliable, relatively simple and clinically useful.