The effect of both training discipline and exercise modality on exercise-induced hypoxaemia (EIH) was examined in seven runners and six cyclists during 5 min high intensity treadmill and cycle exercise. There were no significant interactions between training discipline, exercise modality and arterial P(O(2)) (Pa(O(2))) when subject groups were considered separately but when pooled there were significant differences between exercise modalities. After min 2 of exercise arterial hydrogen ion concentration, minute ventilation, alveolar P(O(2)) (PA(O(2))) and Pa(O(2)) were all lower with treadmill running with the largest differential for the latter occurring at min 5 (treadmill, 80.8+/-1.8; cycle, 90.2+/-2.5, mmHg, N=13, P< or = 0.05). At every min of exercise, the differences in Pa(O(2)) between the ergometers were strongly associated with similar differences in PA(O(2)) and alveolar to arterial P(O(2)) (PA(O(2))-Pa(O(2))). It is concluded that the greater EIH with treadmill running is a consequence of the combined effect of a reduced lactic acidosis-induced hyperventilation and greater ventilation-perfusion inequality with this exercise mode.