1. Two-hundred and forty duplicate estimations of physiological dead-space volume (VD) were made in forty-eight healthy subjects (twenty-four men and twenty-four women) aged from 20 to 74 years, to assess the predictive accuracy of various standards.
2. The VD/VT (physiological dead-space volume/tidal volume) ratio standard was least precise, but could be improved by allowing for sex and age.
3. The best prediction could be made by multiple regression of VD on age, height, tidal volume (VT) and the reciprocal of respiratory frequency (f), which gave an estimate with a standard deviation of 24·7 ml.
4. Theoretical and practical arguments favour the abandonment of the VD/VT ratio standard. Simple regression of VD on VT also is unsatisfactory, giving a much less precise estimate of VD than a multiple regression on VT and other variables.