The factors determining maximal oxygen consumption were explored in eight endurance trained subjects (TS) and eight untrained subjects (US) exposed to moderate acute normobaric hypoxia. Subjects performed maximal incremental tests at sea level and simulated altitudes (1,000, 2,500, 4,500 m). Heart rate (HR), stroke volume (SV), cardiac output (.Q), arterialized oxygen saturation (Sa'O2), oxygen uptake (.VO2max), ventilation (.VE, expressed in normobaric conditions) were measured. At maximal exercise, ventilatory equivalent (.VE/.VO2max), O2 transport (.QaO2max) and O2 extraction (O2ERmax) were calculated. In TS, .Qmax remained unchanged despite a significant reduction in HRmax at 4,500 m. SVmax remained unchanged. .VEmax decreased in TS at 4,500 m, .VE/.VO2max was lower in TS and greater at 4,500 m vs. sea level in both groups. Sa'O2max decreased at and above 1,000 m in TS and 2,500 m in US, O2ERmax increased at 4,500 m in both groups. .QaO2max decreased with altitude and was greater in TS than US up to 2,500 m but not at 4,500 m. .VO2max decreased with altitude but the decrement (Delta.VO2max) was larger in TS at 4,500 m. In both groups Delta.VO2max in moderate hypoxia was correlated with Delta.QaO2max. Several differences between the two groups are probably responsible for the greater Delta.VO2max in TS at 4,500 m : (1) the relative hypoventilation in TS as shown by the decrement in .VEmax at 4,500 m (2) the greater.QaO2max decrement in TS due to a lower Sa'O2max and unchanged .Qmax 3) the smaller increase in O2ERmax in TS, insufficient to compensate the decrease in .QaO2max.