The significance of torsional loads on the tibial component of total knee replacement prostheses is not clearly established. This study investigated the micromotion induced by torsion, under physiologic loads. Three forms of fixation were sequentially tested in cadaveric tibiae: press-fit, press-fit plus 4 peripheral corticocancellous screws, and horizontally cemented. The implanted tibia was loaded with an axial force between 0.5 and 2.2 kN and a cyclic torque of +/- 5 Nm. The relative motion between the prosthesis and the bone was measured. Our results show little difference between the 3 methods of fixation at high axial loads. At low axial loads, however, the press-fit prosthesis may permit micromotions high enough to interfere (theoretically) with bone ingrowth. Motion is reduced when the same prosthesis is augmented with screws. The cemented prosthesis produces the lowest average micromotion. Our study indicates that rotational micromotion is unlikely to contribute to the failure of ingrowth into the uncemented tibial prosthesis under normal physiologic loads. At low axial loads, which may be encountered in the immediate postoperative period, ingrowth may be compromised. Thus, initial rotational stability is an important consideration in uncemented total knee replacement design and postoperative management. In poor-quality tibiae, cementation may be the only method of fixation to provide sufficient torsional stability.