Knee bolsters on the lower instrument panel have been designed to control occupant kinematics during sudden deceleration. However, a wide variability in car occupant anthropometry and choice of seating posture indicates that lower-extremity contacts with the impingement bolster could predominantly load the flexed leg through the knee (acting through the femur) or through the tibia (acting through the knee joint). Potential injuries associated with these types of primary loading may vary significantly and an understanding of potential trauma mechanisms is important for proper occupant restraint. Impacts of the bolster panel against the knee or lower leg were simulated in 10 human cadaver and anthropomorphic dummy tests and the following aspects were assessed: 1) biomechanical response for lower-extremity impacts, 2) potential mechanisms of skeletal and ligamentous trauma, 3) differences between human cadavers and an anthropomorphic test dummy response, and 4) knee-joint ligament failure characteristics in isolated knee-joint tests.Knee impacts with a 55.9 kg bolster covered mass at 6.0 m/s resulted in frequent avulsion fractures of the posterior cruciate ligament at its osseous attachment to the tibia with peak contact loads of 7.02 kN (7.74 kN peak dummy femur load). In this study, analysis of high-speed movies and radiographs indicated that the bolster loaded a against the tibial tuberosity early in the event, translated the tibia posteriorly and resulted in a stretching of the posterior cruciate ligament. Lower-leg impact produced tibial/fibular fractures or knee-joint ligament failures with peak bolster contact loads of 5.15 kN (4.21 kN peak dummy femur load). Isolated knee-joint tests indicated complete failure of the ligament after 2.26 cm of relative posterior tibial subluxation and a resistive load of 2.48 kN. However, the absolute values of the maximally tolerated loads may be significantly influenced by the deficiencies of the cadaver model and cannot be directly extrapolated for real-life situations. Since the lower extremity of the dummy cannot accomodate translatory motion at the knee joint and the skeletal mass of the dummy significantly exceeds that of the human, substantial kinematic and biomechanical response differences occurred between tested human cadavers and an anthropomorphic dummy.