The onset latency and discharge amplitude of preprogrammed postural responses were evaluated in order to determine if the structure of synergistic activation could be altered by ligamentous laxity at the knee joint. Twelve subjects with unilateral and one subject with bilateral anterior cruciate ligament (ACL) insufficiency were tested while standing on a moveable platform. External balance perturbations (6 cm anterior or posterior horizontal displacements of the platform) were presented at velocities ranging from 15 to 35 cm/s. Perturbations were presented under the following experimental conditions: unilateral and bilateral stance, knees fully straight or flexed, and with ankle motion restricted or free. These stance, knee position, and ankle motion conditions were introduced to alter the stress transmitted to the knee joint during movement of the support surface. The automatic postural response was recorded from the tibialis anterior (T), quadriceps (Q), and medial hamstrings muscles (H) bilaterally. The normal response to an externally induced backward sway involved the automatic activation of T and Q at latencies of 80 ms and 90 ms respectively. Activation of the hamstrings in the non-injured extremity was not coupled with the postural response. Hamstrings are not typically involved in the correction posterior sway because H activation would tend to pull the center of mass further backwards. However, when the response in the ACL-deficient extremity was compared to the non-injured limb: (1) the automatic postural response in the ACL-deficient extremity was restructured to include hamstrings activation (100 ms latency), (2) H activation time was faster and less variable in the ACL-deficient limb, and (3) the ratio of H/Q discharge amplitude integrated over 100 ms and 200 ms from the onset of EMG activation showed a dominance of hamstring activity during unilateral stance on the lax limb. In addition, H/Q ratios integrated over 200 ms showed dominant hamstring activity in the ACL-deficient limb during bilateral stance. (4) Cross-limb comparisons showed greater normalized IEMG amplitudes for T, H, and Q during unilateral stance on the lax limb. These results suggest that a capsular-hamstring reflex is integrated into the existing structure of a preprogrammed postural synergy in order to compensate for ligamentous laxity. Furthermore, the generalized increase of response gain observed during perturbations of unilateral stance on the lax limb indicates that joint afference can modulate central programming to control localized joint hypermobility. A concept of postural control is discussed with respect to the capsular reflex, joint loading and displacement of the center of gravity.