We hypothesized that the reduction of O2-carrying capacity caused by the withdrawal of approximately 450 ml blood would result in slower phase II O2 uptake (VO2) kinetics, a lower VO2peak and a reduced time to exhaustion during severe-intensity cycle exercise. Eleven healthy subjects (mean +/- S.D. age 23 +/- 6 years, body mass 77.2 +/- 11.0 kg) completed 'step' exercise tests from unloaded cycling to a severe-intensity work rate (80% of the difference between the predetermined gas exchange threshold and the VO2peak) on two occasions before, and 24 h following, the voluntary donation of approximately 450 ml blood. Oxygen uptake was measured breath-by-breath, and VO2 kinetics estimated using non-linear regression techniques. The blood withdrawal resulted in a significant reduction in haemoglobin concentration (pre: 15.4 +/- 0.9 versus post: 14.7 +/- 1.3 g dl(-1); 95% confidence limits (CL): -0.04, -1.38) and haematocrit (pre: 44 +/- 2 versus post: 41 +/- 3%; 95% CL: -1.3, -5.1). Compared to the control condition, blood withdrawal resulted in significant reductions in VO2peak (pre: 3.79 +/- 0.64 versus post: 3.64 +/- 0.61 l min(-1); 95% CL: -0.04, - 0.27) and time to exhaustion (pre: 375 +/- 129 versus post: 321 +/- 99 s; 95% CL: -24, -85). However, the kinetic parameters of the fundamental VO2 response, including the phase II time constant (pre: 29 +/- 8 versus post: 30 +/- 6 s; 95% CL: 5, -3), were not altered by blood withdrawal. The magnitude of the VO2 slow component was significantly reduced following blood donation owing to the lower VO2peak attained. We conclude that a reduction in blood O2-carrying capacity, achieved through the withdrawal of approximately 450 ml blood, results in a significant reduction in VO2peak and exercise tolerance but has no effect on the fundamental phase of the VO2 on-kinetics during severe-intensity exercise.