We hypothesized that exercise would cause greater severity and incidence of acute mountain sickness (AMS) in the early hours of exposure to altitude. After passive ascent to simulated high altitude in a decompression chamber [barometric pressure = 429 Torr, approximately 4,800 m (J. B. West, J. Appl. Physiol. 81: 1850-1854, 1996)], seven men exercised (Ex) at 50% of their altitude-specific maximal workload four times for 30 min in the first 6 h of a 10-h exposure. On another day they completed the same protocol but were sedentary (Sed). Measurements included an AMS symptom score, resting minute ventilation (VE), pulmonary function, arterial oxygen saturation (Sa(O(2))), fluid input, and urine volume. Symptoms of AMS were worse in Ex than Sed, with peak AMS scores of 4.4 +/- 1.0 and 1.3 +/- 0.4 in Ex and Sed, respectively (P < 0.01); but resting VE and Sa(O(2)) were not different between trials. However, Sa(O(2)) during the exercise bouts in Ex was at 76.3 +/- 1.7%, lower than during either Sed or at rest in Ex (81.4 +/- 1.8 and 82.2 +/- 2.6%, respectively, P < 0.01). Fluid intake-urine volume shifted to slightly positive values in Ex at 3-6 h (P = 0.06). The mechanism(s) responsible for the rise in severity and incidence of AMS in Ex may be sought in the observed exercise-induced exaggeration of arterial hypoxemia, in the minor fluid shift, or in a combination of these factors.