This study examines the effects of prolonged hypoxia, with and without control of end-tidal CO2 partial pressure (P-ET,P-CO2), on the intensity-weighted mean velocity of blood flow in the middle cerebral artery (<(V)over bar (IWM)>) and on heart rate (HR). Specifically, the time course of the responses, their reversibility with brief periods of hyperoxia and the recovery phase following prolonged hypoxia were air investigated. Twelve subjects were studied, of whom nine provided satisfactory data. A purpose-built chamber was used for the prolonged control of the end-tidal gases, and an end-tidal forcing system was used for generating the brief variations in end-tidal gases. Three 16 h protocols were employed: (1) 8 h eucapnic (average P-ET,P-CO2 = 39 mmHg) hypoxia (end-tidal O-2 partial pressure, P-ET,P-O2 = 55 mmHg) followed by 8 h eucapnic euoxia (P-ET,P-O2 = 100 mmHg); (2) 8 h poikilocapnic (average P-ET,P-CO2 4 mmHg below eucapnia) hypoxia (P-ET,P-O2 = 55 mmHg) followed by 8 h poikilocapnic euoxia (P-ET,P-O2 = 100 mmHg); and (3) control (air inspired throughout). <(V)over bar (IWM)> (using Doppler ultrasound) and HR were measured during brief exposures to hypoxic/euoxic and hyperoxic conditions with P-ET,P-CO2 held 1-2 mmHg above eucapnia, at 0, 20, 240 and 480 min in the first 8 h, and at the same times in the second 8 h. There were no significant trends in <(V)over bar (IWM)> under hypoxic conditions for either hypoxic protocol (ANOVA) and no significant differences between the three protocols for <(V)over bar (IWM)> in hyperoxia (ANOVA). In contrast to <(V)over bar (IWM)> there was a significant increase in HR over time during both hypoxic exposures (P < 0.01, ANOVA). HR increased to a similar extent for the two types of hypoxia, and there was some suggestion that HR remained elevated after the relief of hypoxia. The results suggest that, with the level of hypoxia employed, progressive changes in HR occur, but that this lever and duration of hypoxia has little sustained effect on <(V)over bar (IWM)>.