More than 1,300 Austin Moore hemiarthroplasties have been reviewed in the literature, with no reports of fracture of the stem. Many patients with these hip implants had good function. The lack of stem fractures in patients with good functions has not been explained and contrasts with stem fractures that have occurred in patients with cemented prostheses of other designs during the same time. We used three-dimensional finite-element analysis and free-body diagrams to explain the lack of fractures for this device by a description of the probable load-transfer mechanisms between the prosthesis and the bone. Results from our finite-element analysis indicate that, with good calcar-collar support, the stresses in the stem are small because the stem portion of the prosthesis and the bone are uncoupled and, consequently, do not share the resultant bending moment of the head and abductor forces. If the stem is coupled to the bone so that the resultant bending moment is shared, high stresses in the stem are predicted; such stresses are inconsistent with the complete absence of fractures of these prostheses. The results of the finite-element analysis further showed that loss of calcar-collar support with proximal fixation through the fenestrations resulted in high stresses in the stem and stress shielding of the proximal medial cortex. The uncoupled prosthesis also may be modeled with a free-body diagram as a three-force member loaded at the head, stem tip, and in the proximal region. With this model, it can be shown that the reaction force of the stem tip, and thus the peak bending stress in the stem, increases as calcar-collar support is decreased. If there is no calcar-collar support, proximal support must be provided by some combination of integration of bone in the fenestrations and wedging due to the lateral-medial taper of the device. Stresses in the stem are largest when there is no wedging, but high stresses develop in the cancellous bone in the fenestrations. When there is wedging, stresses in the stem can be low, but stresses in the supporting cancellous bone can be high; additional proximal support through the fenestrations substantially reduces these bone stresses. If reduced stresses in the cancellous bone are indicative of a stable device, these mechanisms indicate that fractures of the Austin Moore prosthesis have not occurred in normally loaded hips because load was transferred primarily either through the collar or by wedging, with additional support at the fenestrations.