The aim of this study was to test at maximal exercise the hypothesis of the temperature-dependence of the kinetics of O2 consumption (VO2), which predicts a greater O2 deficit as muscle temperature is decreased. Six male subjects underwent 3 min exercise bouts at the minimum power eliciting maximum O2 consumption (VO2max), at normal temperature (A) and after cooling the thigh muscles by water immersion (C). Breath-by-breath VO2 was measured together with muscle blood flow (Qm), blood lactate accumulation ("early lactate", eLa), heart rate and muscle temperature (Tm). The O2 deficit was calculated by standard procedure. Net VO2max was 2.92 +/- 0.85 (SD) and 3.19 +/- 0.71 l center dot min-1 in C and A respectively (P < 0.05). Correspondingly, maximum power was 20 W lower in C than in A. At exercise start, Tm was 35.0 +/- 1.2 and 27.5 +/- 1.8 degrees C in A and C respectively. O2 deficit was 2.25 +/- 0.53 and 3.05 +/- 1.12 l in A and C respectively. The corresponding eLa was 7.7 +/- 2.5 and 13.8 +/- 2.5 mM, (P < 0.05) while Qm was 376 +/- 92 and 290 +/- 50 ml center dot kg-1 center dot min-1 (P < 0.05) in A and C, respectively. The eLa increase in C is associated with an impaired muscle blood flow and decreased muscle O2 unloading, and does not completely explain the greater O2 deficit in C. The unexplained fraction of the latter is perhaps accounted for by a greater net alactic O2 deficit, in agreement with a temperature-dependent decrease of the velocity constants of oxidative reactions, as suggested by the tested hypothesis.