Eight male subjects were asked to swim 25 m at maximal velocity while the use of the arm(s) and legs was alternately restricted. Four situations were examined using one arm (1A), two arms (2A), one arm and two legs (1A2L) and both arms and legs (2A2L, normal swim) for propulsion. A significant mean increase of 10% on maximal velocity was obtained in 1A2L and 2A2L compared to 1A and 2A. A non-significant 4% effect was obtained in 1A. This study focused on the actual contribution of leg kick in the 10% gain in maximal velocity. It was clear that the underwater trajectory of the wrist was modified by the action of the legs (most comparisons P < 0.001). Therefore it was thought that the legs enhanced the generated propulsive force by improving the propulsive action of the arm. The arm action was quantified by selecting typical phases from the filmed trajectory of the wrist, namely forward (F), downwards (D) and backwards (B). Although there was a tendency for individual changes in kinematic parameters (F, D and B) to occur with individual changes in velocity when 2A was compared to 2A2L, no relationship was found between the relative changes in F, D and B and relative changes in velocity. This was illustrated by describing the responses of three individuals who could represent three patterns of contribution by legs and arms to propulsion in high speed swimming.