On a mechanical model of the lungs systematic (constant) and accidental (random) errors in measuring the volume by the device (see article)-1, were determined and these amounted to +2.7 and identical to 2.2, respectively. The latter was three times as low as the mean-square reproducibility error, i.e. the degree of error in the physiological measurement of FRC. Different sources of errors in the measurement of FRC, among which particular attention is called to errors in defining the helium concentration at the end of the mixing process in closed system, are considered.