To investigate the effect of exercise mode on arterial oxyhemoglobin saturation (Sa O 2 ), 13 healthy, actively training men who displayed exercise-induced hypoxemia (EIH) performed two incremental maximal exercise tests: uphill treadmill running and cycle ergometry. At maximum, treadmill running resulted in a lower Sa O 2 (88.6+/-2.0% versus 92.6+/-2.0%), a lower ventilatory equivalent for carbon dioxide (V E V C O 2 ; 28.8+/-0.6 versus 31.2+/-0.9), and a higher maximal oxygen consumption (V O 2 , M A X ; 4.83+/-0.11 l.min - 1 versus 4.61+/-0.14 l.min - 1 ) when compared to cycle ergometry. When data were combined from maximal running and cycling, Sa O 2 was correlated to V E V C O 2 (r=0.54). However, there was no relationship between the differences in Sa O 2 and ventilation between exercise modes. This suggests that ventilation is important in the maintenance of Sa O 2 , but that the difference observed in Sa O 2 between treadmill running and cycle ergometry cannot be explained by differences in ventilation and must be due to differences in diffusion limitation or ventilation-perfusion inequality.