The experiment was undertaken in response to the three questions previously posed. We believe that these questions have been satisfactorily if not completely answered:

1. When the ends of a ruptured ligament are in approximation, healing occurs by an orderly process: first, by necrosis and cellular infiltration, then by fibroblastic invasion, and finally, by collagenization. If the ligament ends are not in close approximation, this orderly process does not occur and healing is by diffuse extensive scar formation. As time passes, the scar at the site of rupture becomes more orderly and its structure begins to resemble that of a normal ligament. Up to ten weeks, however, this scar had not become normal ligament even under the most favorable circumstances. Possibly, normal ligament would have developed if the process had been followed longer.

2. The most favorable circumstance for repair is end-to-end direct apposition of ligament fibers. This minimizes scarring, accelerates repair, hastens collagenization, and produces the most normal-appearing healed ligament.

3. Tests of the tensile strength of the divided lateral collateral ligament in our dogs suggest that after six or more weeks of immobilization, the strength of the attachment of the ligament to bone is diminished, presumably as the result of the atrophy of disuse. It would appear that during the first few weeks after such immobilization, care should be taken to prevent damage to this weakened area. This will need further confirmatory work.