The effect of varying the preinspiratory volume (V iair; range: 0–75% vital capacity, VC) on the nitrogen closing volume (CV) test was studied in twelve seated subjects, aged 24–62 years. When V iair was increased from 0 to about 12% VC, the height of phase IV, the amplitude of the cardiogenic oscillations, CV and the slope of phase III increased. The height of phase IV and the amplitude of the cardiogenic oscillations showed a maximum at V iair = 12% VC , although the average CV was about 18% VC. While the height of phase IV and the amplitude of the cardiogenic oscillations decreased when V iair was increased above 12% VC, CV did not change and the slope of phase III increased consistently. These results cannot be explained solely by the regional lung volume model of Sutherland et al. (1968). However, if that model is extended to include the assumption that within a region alveoli behind closed airways may be differently expanded, we predict CV to be underestimated at low V iair, independently of the upper to lower concentration difference, in agreement with present findings. This assumption would also explain why the maximal height of phase IV can be obtained at a V iair lower than CV.