The reproducibility of measurements of blood flow velocity in the aorta and systolic time intervals using pulsed ultrasound Doppler was determined in 10 subjects. Measurements were made in both the ascending aorta and the aortic arch, and instant maximum as well as instant mean velocities were measured. The smallest variation was found when the maximum velocity was measured in the ascending aorta. The coefficient of variation in measurements was between 6.9% and 13.1%, when determined at intervals of 10-15 min. Again the smallest variation was when the maximum velocity was measured in the ascending aorta. The intraobserver variation was smaller than the interobserver variation (6.9% compared with 10.6%) in the ascending aorta. Both the intraobserver and interobserver variations in systolic time intervals were very small. The mean coefficients of variation were 1.7% for the pre-ejection period (PEP), 1.7% for left ventricular ejection time (LVET), 1.4% for electromechanical systole (QS2), and 2.9% for the PEP:LVET ratio. The mean values of pre-ejection period and left ventricular ejection time were 154 ms and 377 ms, which are a little longer and shorter respectively than measurements made by phonocardiography and pulse pressure monitoring, whereas QS2 was the same (531 ms). The interobserver variations were 2.6%, 2.0%, 1.4%, and 3.9% respectively. It is concluded that pulsed Doppler ultrasound is a reliable method of measuring blood flow velocity and systolic time intervals. Changes in blood flow velocity of 10-12% can be detected by this method.