A crossover design was used to evaluate kinematic measurements collected with an infrared-based motion measurement system. To evaluate belt effects on spine kinematics during asymmetric lifting of large and small boxes and to test for carryover effects between trials from belts. Conflicting evidence in the literature exists regarding whether belts are beneficial or detrimental to manual material handlers. Studies have not examined belt effects when lifting different sized boxes, nor carryover effects from belts. Twenty-eight subjects with manual-handling experience (17 male and 11 female) were randomly assigned to lift either a large or small box (weighing 9.4 kg), from a sagittally symmetric origin at pallet height to a 79 cm height, 60 degrees to the right. Spine flexion, lateral bending and twisting, hip and knee flexion, and angular velocity measurements of the torso with respect to the pelvis were collected for each of three lifting periods, 50 lifts each at 3 lifts per minute, with 18-minute breaks between periods. Belts significantly reduced maximum spine flexion, spine flexion and extension angular velocities, and torso left lateral bending angular velocity, and increased hip and knee flexion, regardless of box size. When lifting large boxes, belts significantly reduced torso right lateral bending and torso left twisting. No significant differential carryover effects were detected from belts. Subjects with belts lifted more slowly and used more of a squat-lift technique, regardless of box size. Belts reduced more torso motions while lifting large boxes.