1. The lung nitrogen from ten normal nonsmoking subjects and ten patients with chronic respiratory disease was washed out by inspiration of 21% oxygen and 79% argon, whilst expired nitrogen concentration was measured with a mass spectrometer and flow with a box-bag system, and the quantity of nitrogen in each expirate calculated with a Varian 73 digital computer and plotted against expired volume.
2. The resulting curve from each breath was handled either as a linear or a polynomial regression, the intercept on the abscissa being designated the series dead space volume (VDS). This dead space has also been measured by the Fowler method and by the mathematical differentiation of phase II.
3. The nitrogen recovered from the first breath varied between 190 ml and 584 ml and this has been expressed as a percentage of that volume predicted assuming perfect alveolar mixing; this has been called the alveolar gas mixing efficiency for nitrogen. The mean values for alveolar mixing efficiency computed from the four different series dead space volumes were 89.8%, 90.8%, 89.0% and 85.7%. This suggests that the value of series dead space used in computing mixing efficiency is not of great importance, and any of the four methods gives satisfactory results.
4. The data from multi-breath washout were also similarly expressed as a percentage efficiency, making the results from the two methods directly comparable. The median value for alveolar gas mixing efficiency by the single breath test was 89%, and for the multi-breath test about 76%. The latter was apparently more discriminating for ventilatory defect. In patients the values were respectively 73% and 40%.