This study focused on response corridors used to evaluate the biofidelity of the lower limbs of whole-body pedestrian dummies. Specifically, corridors for the thigh and the leg were investigated. The three-point bending tests of horizontally placed specimens of the thigh and the leg in past studies exhibited sagging flesh caused by gravity. This resulted in unrealistically thin flesh on the loaded side. This study investigated a methodology to eliminate the influence of the sagging flesh to develop more realistic corridors. The initial toe region of the force- deflection response from the experiment was eliminated from the test results to diminish the influence of unrealistic flesh thickness. Due to the difference in stiffness, the flesh would bottom out upon initiation of major bone deflection. The assumption was made that the magnitude of the force applied at the beginning of this phase is independent of flesh thickness. Dynamic three-point bending simulations were conducted using thigh and leg FE models with the flesh thickness varied to validate the assumption. In addition, existing experimental data were re- examined to determine the initiation of major bone deflection, and to calculate the magnitude of applied force at the same timing. Furthermore, full-scale car-pedestrian impact simulations were conducted using a human FE model with various flesh thicknesses to clarify the influence of the initial toe region of the force-deflection response. The corridors were then developed using the data after the applied force reached a specific value determined in this study. The results of three-point bending simulations with various flesh thicknesses showed that the magnitude of applied force at the initiation of major bone deflection was fairly constant, at approximately 2000 N and 1500 N for the thigh and the leg, respectively. These values were found to eliminate the initial toe region from the experimental data, while maintaining the stiffness of the subsequent region that is not influenced by the initial thickness of the flesh. The full-scale impact simulations showed that the peak values of the femur and tibia bending moments were not significantly influenced by the flesh thickness.