To test the hypothesis that continuous fetal breathing could be induced by hyperoxemia alone or by hyperoxemia and umbilical cord occlusion, even in the absence of a rise in arterial PCO2 (Pa(CO2)), we studied 18 chronically instrumented fetal sheep on 34 occasions using our window model. After a resting cycle (1 low-voltage followed by 1 high-voltage electrocortical activity epoch), the fetal lung was distended via an endotracheal tube using mean airway pressure of ~ 30 cmH2O. Inspired N2, 17% O2, and 100% O2 were given to the fetus during one cycle each. While 100% O2 was given, the umbilical cord was occluded (balloon cuff). We found that 1) N2 and 17% O2 did not affect breathing and behavior; 2) 100% O2 induced continuous breathing in 12 of 34 experiments (35%), with an increase in arterial PO2 (Pa(O2)) (23 ± 1 to 251 ± 26) and Pa(CO2) (43 ± 1 to 51 ± 2 Torr) and a decrease in pH (7.39 ± 0.01 to 7.32 ± 0.01); 3) in 17 experiments 100% O2 did not induce continuous breathing, with the rise in Pa(O2) being modest (to 103 ± 25 Torr), but subsequent cord occlusion did, with the rise in Pa(O2) being more pronounced (to 190 ± 28 Torr); 4) in 10 experiments, continuous breathing persisted despite keeping Pa(O2) within 10 Torr of control values by use of high-frequency oscillatory ventilation (in 3 of these experiments Pa(CO2) did not change throughout the study; 5) in 5 experiments in which the endotracheal tube was partially obstructed, neither O2 nor cord occlusion induced continuous breathing despite a major increase in Pa(CO2), decrease in pH, and unchanged Pa(O2); and 6) during 100% O2 and cord occlusion the fetus was aroused, with increased movements and occasional opening of the eyes. The findings suggest that low fetal Pa(O2) inhibits breathing and that an increase in Pa(O2) is required for maintenance of continuous breathing in the fetus. We speculate that this increase is important for establishment of continuous breathing at birth.