The cause of headache in persons going to high altitude is unknown. Relatively severe hypoxemia in susceptible subjects could induce large increases in cerebral blood flow that then could initiate the headache. Thus we measured noninvasively, by Doppler ultrasound, changes in internal carotid arterial blood velocity (velocity) in 12 subjects in Denver (1,600 m) and repeatedly up to 7 h at a simulated altitude of 4,800 m (barometric pressure = 430 Torr). Six subjects, selected because of prior history of high-altitude headache, developed comparatively severe headache at 4,800 m, and four subjects, without such history, remained well. Two subjects developed moderate headache. Velocity at 4,800 m did not correlate with symptom development, arterial O2 saturation, or end-tidal PCO2. Also, neither velocity nor blood pressure was consistently elevated above the Denver base-line values. During measurements of hypercapnic ventilatory response in Denver, velocity increased linearly with end-tidal PCO2, confirming that our Doppler method could demonstrate an increase. Also, 30 min of isocapnic or poikilocapnic hypoxia caused small increases in velocity (+8 and +6%) during the base-line measurement at low altitude. Although even a small increase in cerebral perfusion could contribute to headache symptoms at high altitude, cerebral blood flow does not appear to play a primary role.