The distribution of strain in the proximal part of loaded cadaver femora was measured in vitro using strain gauges applied to the cortex. The loading conditions simulated single-limb stance and the strains were recorded first with the femora intact and then with the femoral components of six different designs inserted. Each femur served as its own control. After insertion of a femoral component, the pattern of strain in the proximal part of the femur was reversed compared with that in the intact femur, in that the maximum strain occurred around the tip of the prosthesis rather than at the calcar femorale. A massive decrease in stress in the region of the calcar femorale was found when the implants were in place, and it was concluded that this decrease could contribute substantially to the calcar femorale resorption sometimes observed in patients after total hip replacement. Transfer of load directly to the calcar femorale through a larger collar in direct contact with the cortical bone restored 30 to 40 per cent of the normal strain to the calcar femorale and shifted the strain pattern toward normal. Compared with the less stiff stems tested, the larger, stiffer stems, which provide more protection against fatigue failure, did not affect the strain pattern adversely.