Two kinds of studies have been conducted in order to measure respiratory water loss: a single breath study of instantaneous variations in relative gas humidity of air expired during one respiratory cycle and a multibreath study of the average values of water vapor in air expired during several successive cycles of steady state ventilation. In the first case, relative gas humidity is computed from results obtained by thermometry and mass spectrometry; in the second case, average water vapor content of expired air is calculated from plethysmographic spirometry and expired water collection. Both experiments showed that mixed expired gas is not fully water saturated. The multibreath study showed that the mass of water lost per liter of ventilated gas is not a function of ventilation per se but rather increases as tidal volume rises and decreases as respiratory frequency diminishes. The mass of water lost per cycle of steady state ventilation increases with tidal volume so that mean expired gas volume may be considered as a mixture of dry gas and water saturated gas. The single breath study showed that unsaturated gas is expired in the first part of expirate followed by wet saturated gas in the second part. The numerical values given by the two kinds of studies are in close agreement.