A four-fragment trochanteric fracture was produced in 24 human cadaver femora. After reduction to anatomic position, the fractures were stabilized by either a Jewett 135 degree nail-plate, a 135 degree Nolok sliding screw-plate without a key, or a 140 degree Hansson pin-plate with an additional trochanteric plate. Using a hip-force simulator, the preparations were subjected to 10,000 loading cycles simulating partial weight bearing, followed by another 10,000 cycles with full weight bearing, while measuring the elastic and permanent fracture displacements. Failure occurred in 3/8 of the Jewett stabilized fractures and in 1/8 each of the Nolok and Hansson stabilized specimens. Each failure was preceded by an increase in elastic tilt and rotation followed by a permanent rotation after which failure occurred as varus angulation of the fracture. Further, two Jewett nails, one Nolok screw, and three Hansson pins showed bending. Due to the controlled telescopic action in the Nolok and Hansson dynamic devices, the resistance to fatigue when subjected to repetitive loadings was increased and thereby a better overall result was achieved when compared with the rigid Jewett nail-plate. To maintain the dynamic function, the strength of the dynamic screw-pin must be sufficient to withstand physiologic loadings without bending; otherwise, the telescopic action will be obstructed.