The mechanism by which flagella generate the propulsive force for movement of hamster spermatozoa was analyzed quantitatively. Tracing points positioned 30, 60, 90, and 120 μm from the head-midpiece junction on the flagellum revealed that they all had zigzag trajectories. These points departed from and returned to the line that crossed the direction of progression. They moved along the concave side (but not the convex side) of the flagellar envelope that was drawn by tracing the trajectory of the entire flagellum. To clarify this asymmetry, the bending rate was analyzed by measuring the curvatures of points 30, 60, 90, and 120 μm from the head-midpiece junction. The bending rate was not constant through the cycle of flagellar bending. The rate was higher when bending was in the direction described by the curve of the hook-shaped head (defined as a principal bend [P-bend]) to the opposite side (R-bend). We measured a lower bending rate in the principal direction (R-bend to P-bend). To identify the point at which the propulsive force is generated efficiently within the cycle of flagellar bending, we calculated the propulsive force generated at each point on the flagellum. The value of the propulsive force was positive whenever the flagellum bent from an R-bend to a P-bend (when the bending rate was lowest). By contrast, the propulsive force value was zero or negative when the flagellum bent in the other direction (when the bending rate was higher). These results indicate that flagellar bending in hamster spermatozoa produces alternate effective and ineffective strokes during propulsion.