Ventilation and Pa(CO2), were measured in six subjects after 10-12 min of breathing 1-2% CO2 during hyperoxia and hypoxia. These inspired CO2 concentrations were achieved in two ways: by enriching the inspirate with CO2 and by having the subjects breathe through dead spaces of 100-400 cm3. Breathing through dead space gave the same results as CO2 enrichment of the inspirate when the effect of the dead spaces on mean inspired CO2 was allowed for. During hyperoxia all subjects demonstrated isocapni hyperpnea in response to mean inspired CO2 concentrations of 1%; ventilation increased without change in PA(CO2). When mean inspired CO2 concentration approximated 1.5% two subjects showed isocapnic hyperpnea, and one subject demonstrated isocapnic hyperpnea in response to mean inspired CO2 concentrations of 2%. The increase in PA(CO2) observed in each subject in response to 2% CO2 in O2 correlated negatively with the slope of that subject's rebreathing CO2 response curve. Hypoxia (PA(O2Y = 45-50 mm Hg) depressed the response to 1% CO2 in that, while hypoxic, no subject showed isocapnic hyperpnea in response to 1% CO2. The isocapnic hyperpnea we observed was chiefly due to increased tidal volume, and was therefore not analogous to the isocapnic hyperpnea observed by others in dogs in response to increases of CO2 in lung gas. When low levels of CO2 produced an increase in PA(CO2) the associated change in ventilation (delta Ve/delta PA(CO2)was much less than that observed while rebreathing 7% CO2. Isocapnic hyperpnea in response to low levels of CO2 is common among normal individuals, and is depressed by hypoxia; the stimulus responsible for this response is unknown.