This study compared cardiac stroke volumes using four noninvasive techniques in healthy volunteers: dual-beam Doppler echoaortography, electrical impedance cardiography, mechanosphygmography and M mode echocardiography of the left ventricle. Variations in stroke volume were induced by intravenous atropine in increasing doses and venous tourniquets in the absence and presence of propanolol. Biometrical analysis used correlations and principal components in order to find the common reality, reflected by the four techniques. Three of the four techniques showed good correlations between themselves as well as to the principal component. The correlation closest to the principal component was the electrical impedance cardiography (0.89) followed by the Doppler method (0.87) and the mechanosphygmography (0.84); the results from the left ventricular echocardiography (0.66) were poorest. These results can be interpreted in the sense that the first three techniques are of good validity in the noninvasive assessment of stroke volume in man, and the selection of one of these techniques for use in a clinical study should depend on the study's design and scope.