A non-invasive technique for the measurement of cardiac output in pregnancy by combined cross-sectional and Doppler echocardiography at three intracardiac sites is described. The validity of the technique for use during pregnancy is reviewed. Comparison with cardiac output determined simultaneously by the direct Fick technique in 15 non-pregnant subjects showed close agreement for all three measurement sites. Acceptable measurements were obtained from the aortic and pulmonary valves in all pregnant subjects and from the mitral valve in 84% of pregnant subjects. The within-patient intra-observer and hour-to-hour variabilities of cardiac output in pregnant and non-pregnant subjects were less than 5% of the mean for all three valves studied. Flow measurements at each of the three intracardiac sites correlated closely. The advantages and limitations of the technique for use during pregnancy are discussed.