To determine ideal alignment and component placement of total knee prostheses, Kinematic (K) and total condylar (TC) devices were physiologically loaded and interface forces were measured. Laboratory observations were correlated with clinical (roentgenographic) findings. Asymmetric loading of the tibial component has been proposed as causing loosening and radiolucent lines. Misalignment of components is one factor that affects load sharing by bone under the medial and lateral regions of the tibial plateau. Tibial components of K and TC prostheses were inserted without cement into the cut surfaces of artificial tibiae. The mating femoral condylar components were mounted. The tibial and femoral components were individually positioned at 0 degrees (horizontal) and at certain angles of varus and valgus. Pressure-sensitive film was placed between the tibial component and the artificial tibia. A vertical load of 1500 N was used. The experiment was replicated twice. The percentages of the load on the medial and lateral regions of the tibial plateau were calculated from quantitative image analysis of the pressure patterns on the film. Roentgenograms from 532 K and 21 TC patients were examined to determine the orientations of the condylar and tibial components and the presence of radiolucent lines around the tibial component. An even distribution (ideal alignment) of load on the medial and lateral regions of the K tibial component occurred at 9 degrees of valgus tilt of the femoral component and 2 degrees of varus tilt of the tibial component and for the TC at 7 degrees valgus and 0 degrees varus. Misalignment by 5 degrees yielded a 7% change in the load distribution under the K plateau and a 40% change for the TC prosthesis; a 10 degrees misalignment produced changes of 34% and 62% for the K and TC, respectively. Small variations in clinical knee alignment produced the same percentage of radiolucent lines for each alignment group. The location of radiolucent lines was distributed among the medial, lateral, and both tibial plateaus regardless of knee alignment, although there were more medial reactions overall. The smallest incidence (8%) of radiolucent lines occurred with the K prosthesis at 7 degrees of knee valgus, the femoral component placed at 9 degrees valgus, and the tibial component at 2 degrees varus. This correlated with the ideal bench-test findings for the K device.