We wished to determine the optimal tension required to restore normal joint laxity to anterior cruciate ligament (ACL)-deficient knees using a braided polyethylene ACL prosthesis (PACL). In 10 cadaveric specimens, we measured the anteroposterior (AP) laxity of the intact knee at 10 degrees, 30 degrees, 60 degrees, and 90 degrees of flexion. The ACL was then removed and replaced with the PACL using tunnel-tunnel (T-T) and "over-the-top" (OTT) placement techniques. In both positions, the PACL was initially tensioned to 0, 9, 18, and 27 N with the knee flexed to 30 degrees. AP joint laxity was then measured at each flexion angle. With an increase in initial tension, there was a corresponding decrease in AP laxity. At 30 degrees and 90 degrees of flexion, AP laxity was not significantly different from normal using T-T placement and an initial tension of 0 N. At 90 degrees of flexion, AP laxity was not significantly different from normal using OTT placement at 0 or 9 N of initial tension. For both positions, all other tension levels and flexion angles constrained AP laxity. No laxity differences were detected between the OTT and T-T positions at any flexion angle. The variability in AP laxity of the T-T position was significantly greater than OTT. With a 150-N anterior shear force applied to the proximal tibia, the maximum tensions developed in the PACL were not significantly different between the two positions except at 90 degrees. The results suggest that implantation of the PACL is best performed using OTT positioning with an initial tension of 0 N applied at 30 degrees of knee flexion.