BACKGROUND: There are many descriptions of medial collateral ligament (MCL) reconstruction, but they may not reproduce the anatomic structures and there is little evidence of their biomechanical performance. PURPOSE: To investigate the ability of "anatomic" MCL reconstruction to restore native stability after grade III MCL plus posteromedial capsule/posterior oblique ligament injuries in vitro. STUDY DESIGN: Controlled laboratory study. METHODS: Twelve cadaveric knees were mounted in a kinematic testing rig to impose tibial displacing loads while the knee was flexed-extended: 88-N anteroposterior translation, 5-N·m internal-external rotation, 8-N·m valgus-varus, and combined anterior translation plus external rotation (anteromedial rotatory instability). Joint motion was measured via optical trackers with the knee intact; after superficial MCL (sMCL), deep MCL (dMCL), and posterior oblique ligament transection; and then after MCL double- and triple-strand reconstructions. Double strands reproduced the sMCL and posterior oblique ligament and triple-strands the sMCL, dMCL, and posterior oblique ligament. The sMCL was placed 5 mm posterior to the epicondyle in the double-strand technique and at the epicondyle in the triple-strand technique. Kinematic changes were examined by repeated measures 2-way analysis of variance with posttesting. RESULTS: Transection of the sMCL, dMCL, and posterior oblique ligament increased valgus rotation (5° mean) and external rotation (9° mean). The double-strand reconstruction controlled valgus in extension but allowed 5° excess valgus in flexion and did not restore external rotation (7° excess). The triple-strand reconstruction restored both external rotation and valgus throughout flexion. CONCLUSION: In a cadaveric model, a triple-strand reconstruction including a dMCL graft restored native external rotation, while a double-strand reconstruction without a dMCL graft did not. A reconstruction with the sMCL graft placed isometrically on the medial epicondyle restored valgus rotation across the arc of knee flexion, whereas a reconstruction with a more posteriorly placed sMCL graft slackened with knee flexion. CLINICAL RELEVANCE: An MCL injury may rupture the anteromedial capsule and dMCL, causing anteromedial rotatory instability. Persistent MCL instability increases the likelihood of ACL graft failure after combined injury. A reconstruction with an anteromedial dMCL graft restored native external rotation, which may help to unload/protect an ACL graft. It is important to locate the sMCL graft isometrically at the femoral epicondyle to restore valgus across flexion.