The increased utilization of total elbow replacements has resulted in a correspondingly increased number of failed implants requiring revision. The most common reason for revision is aseptic loosening of the ulnar component due to polyethylene induced osteolysis. Implant malalignment is thought to be an important cause of bearing wear and implant failure. The ulnar flexion axis can be used to accurately align the ulnar component of the elbow implant; however, the optimal method of determining this axis intra-operatively is unknown. This in vitro study determined the relationship amongst kinematically and anatomically defined ulnar flexion axes in an effort to improve the accuracy of ulnar component positioning. Five different techniques were used to determine the ulnar flexion axis in 12 cadaveric specimens, 3 kinematic and 2 anatomic. The techniques were compared with the screw displacement axis from simulated elbow flexion. An anatomic measurement technique using the guiding ridge of the greater sigmoid notch of the ulna and the radial head was found to most accurately replicate the position and orientation of the screw displacement axis of the elbow . Because an anatomically derived flexion axis can be determined using both pre-operative imaging techniques, as well as with intra-operative guides, it is more practical than kinematically derived techniques requiring tracking systems for clinical application and should provide reliable and consistent results.