We used two protocols to determine if hypoxic ventilatory decline (HVD) involves changes in slope and/or intercept of the isocapnic HVR (hypoxic ventilatory response, expressed as the increase in VI per percentage decrease in SaO2). Isocapnia was defined as 1.5 mmHg above hyperoxic PET(CO2). HVD was recorded in protocol I during two sequential 25 min exposures to isocapnic hypoxia (85 and 75% SaO2, n=7) and in protocol II during 14 min of isocapnic hypoxia (90% SaO2, FIO2=0.13, n=15), extended to 2 h of hypoxia with CO2-uncontrolled in eight subjects. HVR was measured by the step reduction to sequentially lower levels of SaO2 in protocol I and by 3 min steps to 80% SaO2 at 8, 14 and 120 min in protocol II. The intercept of the HVR (VI predicted at SaO2=100%) decreased after 14 and 25 min in both protocols (P<0.05). Changes in slope were observed only in protocol I at SaO2=75%, suggesting that the slope of the HVR is more sensitive to depth than duration of hypoxic exposure. After 2 h of hypoxia the HVR intercept returned toward control value (P<0.05) with still no significant changes in the HVR slope. We conclude that HVD in humans involves a decrease in hyperoxic ventilatory drive that can occur without significant change in slope of the HVR. The partial reversal of the HVD after 2 h of hypoxia may reflect some components of ventilatory acclimatization to hypoxia.