Recent evidence suggested that exercise-induced arterial O2 desaturation may occur in highly trained endurance athletes. So, Dempsey brought the hypothesis that the pulmonary capacity for oxygen transport cannot meet superior demands imposed by cardiovascular system in highly trained endurance athletes, and endurance training primarily causes adaptation in the skeletal muscles and in the systemic cardiovascular system, with little change in the pulmonary system. In the present study, we determined the propriety of the hypothesis due to measure the ventilatory capacity of endurance athletes. Sixteen highly trained endurance athletes (ET) and thirteen untrained subjects (UT) volunteered to participate in these experiments. All subjects performed the four experiments, 1) the highest oxygen uptake (peak VO2) during incremental cycle exercise and ventilation (VE), ventilatory equivalent for O2 (VE/VO2) and arterial O2 saturation (SaO2) at which peak VO2 was observed, 2) the maximal voluntary ventilation for 30 sec at rest (MVV), 3) the pulmonary diffusing capacity for CO (DLCO) and expressed per unit of alveolar lung volume (KCO) at rest by the single breath method, and 4) the ventilatory response to hypercapnia (S) at rest by rebreathing method, were measured. The peak VO2 of ET (66.7 ml.min-1.kg-1) was significantly (30.8%) higher than UT (52.4 ml.min-1.kg-1), and VE/VO2 and SaO2 of ET (29.3 and 93.7%, respectively) were significantly lower than UT (34.6 and 95.8%). There were no differences in VE, MVV, DLCO, and S between two groups.