Persons with acute altitude sickness hypoventilate at high altitude compared with persons without symptoms. We hypothesized that their hypoventilation was due to low initial hypoxic ventilatory responsiveness, combined with subsequent blunting of ventilation by hypocapnia and/or prolonged hypoxia. To test this hypothesis, we compared eight subjects with histories of acute altitude sickness with four subjects who had been asymptomatic during prior altitude exposure. At a simulated altitude of 4,800 m, the eight susceptible subjects developed symptoms of altitude sickness and had lower minute ventilations and higher end-tidal PCO2's than the four asymptomatic subjects. In measurements made prior to altitude exposure, ventilatory responsiveness to acute hypoxia was reduced in symptomatic compared to asymptomatic subjects, both when measured under isocapnic and poikolocapnic (no added CO2) conditions. Diminution of the poikilocapnic relative to the isocapnic hypoxic response was similar in the two groups. Ventilation fell, and end-tidal PCO2 rose in both groups during 30 min of steady-state hypoxia relative to values observed acutely. After 4.5 h at 4,800 m, ventilation was lower than values observed acutely at the same arterial O2 saturation. The reduction in ventilation in relation to the hypoxemia present was greater in symptomatic than in asymptomatic persons. Thus the hypoventilation in symptomatic compared to asymptomatic subjects was attributable both to a lower acute hypoxic response and a subsequent greater blunting of ventilation at high altitude.