In three anesthetized markedly obese patients, non-invasive stroke volumes (the transthoracic electrical impedance method) were compared to simultaneously obtained invasively measured stroke volumes (dye-dilution method). Close correlations were obtained (r = 0.90--0.98) between the two methods, although constant lower impedance stroke volumes were found in these patients when the values usually employed for the electrical resistivity of the blood (zeta) were used for calculation of the impedance stroke volume. No statistically significant difference (P greater than 0.10) between the two methods was found when a zeta of 175 ohm X cm was used for the calculations, or when the percentage changes in stroke volumes were analyzed. The impedance method is safe and reliable; it also permits non-invasive measurements of stroke volume during anesthesia in markedly obese patients.