We tested the hypothesis that kinetics of O2 uptake (V˙o2) measured in the transition to exercise near or above peakV˙o2(V˙o2 peak) would be slower than those for subventilatory threshold exercise. Eight healthy young men exercised at ∼57, ∼96, and ∼125%V˙o2 peak. Data were fit by a two- or three-component exponential model and with a semilogarithmic transformation that tested the difference between required V˙o2 and measuredV˙o2. With the exponential model, phase 2 kinetics appeared to be faster at 125% V˙o2 peak[time constant (τ2) = 16.3 ± 8.8 (SE) s] than at 57%V˙o2 peak(τ2 = 29.4 ± 4.0 s) but were not different from that at 96%V˙o2 peakexercise (τ2 = 22.1 ± 2.1 s).V˙o2 at the completion of phase 2 was 77 and 80%V˙o2 peak in tests predicted to require 96 and 125%V˙o2 peak. WhenV˙o2 kinetics were calculated with the semilogarithmic model, the estimated τ2 at 96%V˙o2 peak (49.7 ± 5.1 s) and 125%V˙o2 peak (40.2 ± 5.1 s) were slower than with the exponential model. These results are consistent with our hypothesis and with a model in which the cardiovascular system is compromised during very heavy exercise.