Osteolysis is the main cause of aseptic loosening and stem failure. The mechanism that leads to osteolysis is poorly understood; pressure generation caused by reversible stem micromotion may play an important role. We aimed to determine whether dynamically inducible micromotion occurs in vivo at the prosthesis-cement interface and to use these data to develop and confirm a finite element representation of this interface. Dynamically inducible micromotion was measured using radiostereometric analysis in 21 hips implanted with an Exeter stem, at 3 months and 12 months postoperatively, by changing loading from double-leg stance to single-leg stance. Dynamically inducible micromotion occurred at 3 and 12 months; similar micromotion was observed at both time points. At 3 months the head of the stem was displaced posteriorly (0.10 +/- 0.16 mm) and inferiorly (0.08 +/- 0.12 mm) on loading. A Coulomb friction nonbonded representation of the stem-cement interface was used to fit the clinically measured dynamically inducible micromotion. The final finite element model predicted gap opening and closing between the implant and the mantle. This may be a mechanism for generating pressure and distributing wear debris, which are believed to important contributors to failure.