This article begins to address the validation requirements of wear testing on total knee replacements in a knee simulator. The knee simulator has four stations. The axial force is variable but reaches a maximum of 2.3 kN. Physiologic anteroposterior shear force and rotational torques are supplied to the knee. The forces and displacements are timed to coincide with those of a typical gait cycle. Kinematics of the simulator are dependent on the type of knee being tested. Tests of designs with well known clinical histories were done to 10 million cycles. The relative amounts and types of wear shown by the designs were similar to that found in their clinical histories. Wear tracks on more conforming designs were larger, and the penetration into the plastic appeared to be less. This did not necessarily mean that wear, as measured by loss of material, was reduced on conforming designs. Delamination of the plastic was achieved only after aging the tibial components. Wear particles isolated from the lubricating fluid were similar in size and shape to those isolated from in vivo specimens. However, the relative amounts of wear particle shapes were different depending on the design. At the start of the tests, all of the flexibly mounted tibial components showed more motion than after 5 million cycles, indicating that the surface of the plastic became more conforming. This study showed that knee wear similar to wear observed in vivo can be reproduced in the laboratory. The parameters and methods elucidated in this introductory study should form the basis for use in preclinical wear tests of total knee replacements.