Collar seating and tight distal stem fit were evaluated for their effects on micromotion of the implant in the proximal femur. Collared and collarless implants were inserted with a cementless technique in 42 preserved femoral specimens from adult cadavers and cyclically axially loaded in a 9 degree valgus position. Micromotion was registered with dial indicators. Micromotion was minimized distally and proximally by achieving tight distal fixation. Collar seating prevented distal migration and failure of fixation, but it did not have a detectable effect on mediolateral or anteroposterior proximal micromotion. Toggle appears to be controlled primarily by the distal portion of the stem. Axial load transfer appears to be primarily controlled by collar seating. The least micromotion was achieved with a tight fit distally.