Aims: There are multiple proximal prosthetic geometries for humeral head replacement for treatment of four-part proximal humerus fractures. We compared four proximal prosthetic geometries in stable and unstable fracture patterns with a standard tuberosity þxation method. Methods: Twelve synthetic shoulders and 4 cadaver shoulders had a simulated four-part fracture created with an oscillating saw. The following proximal prosthetic geometries were used: smooth circular shape (SCS), diamond shape (DS), irregular multiple þn shape (IMFS), and IMFS with deeper þns (IMSDF). A standardized þxation method using vertical sutures, horizontal sutures and medial based cerclage straps was performed. Passive motion from 0–45 degrees was carried out using a robotic articulator at a rate of 10 degrees per second. Interfragmentary displacement was measured from tuberosity to tuberosity as well as tuberosity to the shaft using mercury strain gauges. This was repeated for stable and unstable fracture patterns. Results: When comparing interfragmentary motion between the four different geometries the greatest amount of motion occurred with the SCS in a stable fracture (0.69mm, p< 0.0001) and unstable fracture (0.71 mm, p< 0.0001). The geometry that provided the most stability was the IMFSDF in stable (0.08mm) and unstable (0.09 mm) fracture patterns. Conclusion: The geometry of the prosthetic device does affect the stability of the tuberosity reconstruction. A smooth circular prosthetic design in a stable or unstable fracture pattern does not prevent excessive interfragmentary motion, while an irregular multiple þn shaped prosthesis with deep þns augments the þxation construct even in an unstable fracture pattern.