The purpose of this study was to present a method of determining the contributions made by rotations of the trunk and upper extremity to hand velocity during the front crawl pull, and to illustrate this with an example. Six male swimmers performed front crawl trials at their middle distance pace (1.52 ± 0.12 m.s−1). Their underwater arm stroke was recorded from the front and side using video cameras suspended over periscope systems. Recordings were digitized at 50 Hz and the 3-D coordinates of the upper extremity were obtained using a DLT algorithm. Shoulder kinematics (flexion/extension, transverse abduction/adduction, internal/external rotation) and elbow kinematics (flexion/extension) were then calculated. Trunk roll kinematics were obtained by digitizing above-water video recordings of a fin attached to each swimmer’s back. The contribution of each body segment rotation to hand velocity was computed using |ɷ × r| cos ϕ, where ɷ was the segment’s angular velocity, r was the position vector of the hand from the segment’s axis of rotation, and ϕ was the angle between hand velocity v hand/pool and v (where v = ɷ × r). Analysis revealed that shoulder extension was the joint motion primarily responsible for producing hand velocity during the insweep (relative contribution: min 66% to max 118%). This was due to the angular velocities and hand-to-joint axis distances for shoulder extension being greater than those of the other joint motions analyzed. The other rotations at the shoulder also contributed to hand velocity during the insweep, but to a lesser extent (transverse adduction: 13% to 49%; internal rotation: −1% to +40%). On average, elbow flexion accounted for 25% of the hand velocity in the middle of the insweep. Trunk roll did not make a positive contribution to hand velocity during the insweep phase (–3% to –48%), contradicting the findings of previous studies.