This study examines the potential for a ventilatory drive, independent of mean PCO2, but depending instead upon changes in PCO2 which occur during the respiratory cycle. This responsiveness is referred to here as dynamic ventilatory sensitivity. The normal, spontaneous, respiratory oscillations in alveolar PCO2 have been modified with inspiratory pulses approximating alveolar PCO2 concentrations, both at sea level and at high altitude (c 5000 m, 16400 feet). All tests were conducted with subjects exercising on a cycle ergometer at 60 W. The pulses last about half the inspiratory duration and are timed to arrive in the alveoli during early or late inspiration. Differences in ventilation, which then occur, in the face of similar end-tidal PCO2 values, are taken to result from 'dynamic ventilatory sensitivity'. Highly significant ventilatory responses (early pulse response greater than late) occurred in hypoxia and normoxia at sea level and after more than four days at c 5000 meters. The response at high altitude was eliminated by normalizing PO2 and was reduced or eliminated with acetazolamide. No response was present soon after arrival (<4 days) at base-camp, c 5000 m, on either of two high altitude expeditions (BMEME, 1994 and Kanchenjunga, 1998). The largest responses at c 5000 m were obtained in subjects returning from very high altitude (7100m - 8848m). The present study confirms and extends previous investigations that suggest that alveolar PCO2 oscillations provide a feedback signal for respiratory control, independent of changes in mean PCO2, suggesting that natural PCO2 oscillations drive breathing in exercise.