An increase in ventilation in spontaneously breathing children, when nitrous oxide was withdrawn from the nitrous oxide-halothane inspired mixture, was previously suggested to be a direct effect of decreased alveolar nitrous oxide concentration. To test whether a "gas effect" of oxygen inflow had caused a false increase in tidal volume signal, in the present study nitrous oxide was withdrawn in two stages, with equilibrium at each stage; there was a 7% increase in VT, to 3.88 ml kg-1. During that nitrous oxide withdrawal, despite a constant vaporizer setting for halothane, the alveolar concentration of halothane decreased (from 1.06% to 0.9%). When the alveolar halothane concentration was subsequently deliberately decreased from 1.06% to 0.9% with nitrous oxide maintained constant at 65%, there was a 14% increase in VT, again to 3.88 ml kg-1. It is concluded that nitrous oxide did not produce significant depression of ventilation in children, and that the observed increases in tidal volume were the result of decreases in end-tidal halothane concentration, produced by a decrease in its inspired concentration.