Eucapnic breathing and ventilatory responses to hypercapnia were studied in seven cats during sleep and wakefulness. No significant differences were found in minute ventilation (VE), alveolar ventilation (VA), or alveolar PCO2 (PACO2) between wakefulness (W) and non-rapid-eye-movement (NREM) sleep, but VA and VE were less during rapid-eye-movement (REM) sleep than W, and PACO2 declined during REM compared with NREM. To test the hypercapnic response, cats were required to rebreathe from a bag containing 6% CO2 and 94% O2 (to eliminate the hypoxic response). The response curve was displaced to the right during NREM and REM; the slope was reduced only during REM to a value about 75% of W and NREM. Eye movements, quantifying phasic REM, were only slightly correlated (negatively) with the deviation of ventilation from the response curve. The hypercapnic response was diminished, not eliminated, during REM, even during phasic REM. The reduced slope arose principally from the failure of the expiratory time to shorten with hypercapnia as during W and NREM. The cat's hypercapnic response compared with the dog's, measured by others with the same methodology, suggests that differences between species may be more crucial than methodology in explaining earlier contradictory results.