A proposed modification to the Hybrid III 50th percentile male dummy upper femur appears to reduce the chest response problems resulting from femur-pelvis interaction in test exposures more severe than Standard No. 208 testing. When compared to overall repeatability of tests, the modification did not change other dummy response measurements appreciably. The femur-pelvis interaction problem, referred to as “hip lock”, was thought to occur in certain vehicles when the femurs of a passenger side dummy impacting only an air bag bottomed out against the pelvis structure. If metal-to-metal contact occurred, excessive load could be transferred to the chest, leading to elevated chest responses. The most pertinent signs of hip lock occurring appear to be a large, sharply pointed z chest acceleration, and a distinct positive component of the lumbar spine z force following the main negative component. A Hybrid III dummy with standard femurs was tested in three configurations to provide baseline data: Taurus passenger-side air bag, Taurus passenger-side three-point belt, and Neon passenger-side air bag. A dummy with modified femurs was tested in the same conditions to see if the modified femurs would solve the hip lock problem without drastically affecting other dummy responses. In cases that exhibited hip lock with the standard femurs, the modified femurs appeared to reduce the signs of hip lock by controlling the femur-pelvis interaction and preventing metal-to-metal contact between the upper femurs and the pelvis structure. In the configurations that did not show hip lock with the standard femurs, no major changes in test data occurred.