Although pre-cooling is known to enhance exercise performance, the optimal cooling intensity is unknown. We hypothesized that mild cooling opposed to strong cooling circumvents skin vasoconstriction and thermogenesis, and thus improves cooling efficiency reflected in improved time to exhaustion. Eight males undertook three randomized trials, consisting of a pre-cooling and an exercise session. During the pre-cooling, performed in a room of 24.6 +/- 0.4 degrees C and 24 +/- 6% relative humidity, participants received either 45 min of mild cooling using an evaporative cooling shirt or strong cooling using an ice-vest. A no-cooling condition was added as a control. Subsequent cycling exercise was performed at 65%[Vdot]O(2peak) in a climatic chamber of 29.3 +/- 0.2 degrees C and 80 +/- 3% relative humidity. During the pre-cooling session, mild and strong cooling decreased the skin blood flow compared with the control. However, no differences were observed between mild and strong cooling. No thermogenesis was observed in any conditions investigated. The reduction of body heat content after pre-cooling was two times larger with strong cooling (39.5 +/- 8.4 W . m(-2)) than mild cooling (21.2 +/- 5.1 W . m(-2)). This resulted in the greatest improvement in time to exhaustion with strong cooling. We conclude that the cooling intensities investigated had a similar effect on cooling efficiency (vasoconstriction and thermogenesis) and that the improved performance after strong cooling is attributable to the greater decrease in body heat content.