Sled tests were conducted to investigate the dynamics of a Part 572 dummy as a sfunction of the belt restraint configuration and impact direction. The tests involved a 35 km/h velocity change and 10 g deceleration. An “opened” fixture, free of intervening surfaces, was oriented from frontal (0°), through oblique (±30°,±45°, ±60°), to full lateral (±90°).Restraint by only a lap belt resulted in the dummy's upper body rotating about the lap belt and continuing in the direction of sled deceleration. Restraint by a lap-shoulder belt greatly reduced upper-body displacement. However, the displacement and body loading were strongly dependent on the direction of deceleration, i.e., the orientation of the belt relative to the impact direction.When the belted shoulder was opposite the impact (0° to +90°), the belt retained the upper body for impact angles of 0° to 45°. Although the upper body escaped from the shoulder belt from 60° to 90°, significant kinetic energy was removed from the upper body before escape, even for full lateral deceleration.When the belted shoulder was on the impact side (0° to -90°), the upper body was restrained for all impact angles. However, the shoulder belt acted directly on the neck with increasing load as the impact became more lateral. Addition of lateral torso restraint, such as with a winged seat greatly reduced the loading of the neck by the shoulder belt for all impact angles.