A two-component dynamic model was used to describe the ventilatory response to sustained hypoxia in humans. One component (Xs) represents the stimulating effects of hypoxia and the other component (Xd), the hypoxic ventilatory decline. The total ventilatory response to hypoxia is represented by the sum of the two components. A nonlinearity is included to account for the nonlinear steady-state ventilatory response to hypoxia. A sensitivity analysis of the model indicates that, with a step change in F(ETO2) as the input, all the parameters can be estimated from the data except for the nonlinearity. The relative sensitivity of the parameters from the model analysis was confirmed in an experimental study. However, comparing steps into hypoxia versus steps out of hypoxia we found a decrease in the gains of both components. The most likely explanation for the decrease in the gains is that the combination of Xs and Xd is not entirely additive. Other models may be required to completely describe the ventilatory response to inputs more complex than steps.