A critical analysis of the thermodilution method of cardiac output determination has been presented. The factors determining cardiac output calculations have been discussed and specific sources of error have been analyzed. This method offers advantages of internal electrical calibration, use of physiologic fluid as indicator, minimal recirculation, lack of requirement for a withdrawal system, and a short interval between successive determinations. The method has compared favorably to other methods of cardiac output determination and is easily used in the modern clinical setting because of its technical simplicity and the availability of flow-directed balloon-tipped catheters with thermistors. Specific sources of error include factors influencing the temperature of the injectate bolus, the determination of initial baseline blood temperature, the integration of the time-temperature curve, the technique of injection, and the method of sampling. These errors depend both upon experimental technique and upon the internal calibrations and methods of curve integration as determined by the manufacturer. In general, the method can be used to accurately determine cardiac output when it is applied systematically and the sources of error are understood. © 1979.