We have tested the hypothesis that severe lypoxia causes apnea, regardless of the arterial CO2 and pH, and that extreme hypoxia causes gasping. Acute experiments with airway occlusion and with low inspired oxygen (FIo2) were performed on anesthetized adult dogs and monkeys. Arterial oxygen saturation was recorded continuously with fiberoptic oximetry, and Pco2 by an electrode catheter. In addition, blood samples were obtained for Po2, Pco2, and pH. Apnea was induced regularly when the Pao2 fell below 10 torr, whether the Paco2 was high with asphyxia (63 torr) or low (26 torr) with low FIo2. Similarly, the Pao2 at apnea was the same whether the pH was 7.17 with asphyxic hypoxia or 7.46 with hypoxic hypoxia. Gasping occurred at even lower Pao2 (below 5 torr) after 1 or 2 min of apnea. Gasping promptly restored the Pao2 to levels of moderate hypoxia (over 30 torr) which permitted resumption of regular respiration, with gradual elimination of the gasping. Fetal monkeys at term were studied in a similar manner from the moment of cord clamping. Their blood gases with apnea were quite similar to adult values in the narrow range of Pao2 and the wide range of Paco2 and pH. In the fetus, gasping was less immediately effective in improving arterial oxygen, but more persistent than in the adult. Regular respirations would not develop in the absence of oxygen in either the fetus or adult animal.