An existing goat model was used to measure in vivo graft forces during walking, to determine if the forces set at surgery change over time under the same external load and if the forces in the graft during in vivo function can be dictated by the forces set at surgery. The anterior cruciate ligament was reconstructed in 12 goats with use of a composite graft consisting of a bone-patellar tendon-bone autograft and a synthetic augmentation segment. The forces in the graft segments were established intraoperatively by a force-setting technique. In five animals, the tendon segment was set to carry 90% of the total graft force, and in the seven other animals, the augmentation segment was set to carry 90% of the total force. The total graft force was the same in all animals. Graft forces due to anterior tibial loads of 67 N were measured before and after fixation and 6 weeks after surgery with the use of buckle transducers mounted extra-articularly over the anterior tibia. They were also measured during straight, level walking at 6 weeks. The forces changed significantly from just after surgery to 6 weeks later, such that the initially set load-sharing was eliminated by 6 weeks. At 6 weeks, a relatively smooth gait had been achieved, and the maximum total graft force in each animal during walking averaged 35 N and was of similar magnitude to forces generated by the anterior tibial loads of 67 N with the animal anesthetized. After fixation, forces in the tendon graft segments were significantly different between the group with high set forces and that with low set forces. At 6 weeks, when functional joint loads were approaching normal levels, the graft segment forces for the two groups were not significantly different. Load-sharing between tendon and augmentation segment and load in the tendon segment at 6 weeks could not be dictated at surgery.