The ventilatory response (VR) of the full-term newborn to CO2 was studied during quiet sleep using rebreathing and steady-state methods. Rebreathing responses were obtained under normoxic (fractional inspired oxygen concentration, Fi,O2 = 0.21) and hyperoxic (Fi,O2 = 0.4) conditions. Ten infants were tested three to five times using each of the two rebreathing protocols and the results averaged. Overall, there was no significant difference between either the mean variability (coefficient of variation) or slope of the VR measured under normoxic and hyperoxic conditions. Four infants were studied using a steady-state technique. There was marked test-to-test variability in VRs measured by this method. The results appear to indicate that the variability of the VR of the newborn to CO2 is not a result of chemoreflex changes in ventilation elicited in response to fluctuations in arterial PO2.