There are many potential applications for cardiac output measurement in clinical and experimental medicine. The most commonly used techniques are invasive procedures, requiring cardiac or arterial catheterization, a disadvantage that has restricted their wider application. Impedance plethysmography has been developed as a non-invasive, beat-by-beat method of cardiac output measurement, which provides an estimate of stroke volume from changes in the electrical impedance of the thorax during cardiac systole. The values for cardiac output obtained by this technique have been extensively compared with values obtained by other methods, both in experimental animals and in the human subject. In the majority of studies high correlation coefficients have been obtained, although impedance plethysmography has tended to give higher absolute values than most other methods. Values for cardiac output obtained by impedance plethysmography are best assessed by comparison with a series of normal values obtained by this technique, rather than with values obtained by other methods. We discuss the results of an investigation of normal cardiac output by impedance plethysmography; theoretical objections to impedance cardiography are considered, and various methods of determining the specific resistivity of blood are reviewed.