Impedance cardiography was used to estimate cardiac output in 10 men during rest and within 5 s after exercise on a bicycle ergometer, including work up to and including maximal aerobic capacity. An indwelling venous catheter permitted simultaneous sampling of venous blood for observing changes in hematocrit associated with each exercise level. Cardiac output, calculated from a standard equation which assumes a constant value of 150 omega.cm for the electrical resistivity of blood, was compared with corresponding calculations in which blood resistivity was individually determined as a function of hematocrit. It is concluded that many of the discrepancies in the literature related to values for cardiac output obtained during exercise by the impedance method may be inherent in calculations that do not consider the changing electrical resistivity of the blood with a changing hematocrit.