The use of projectional techniques to evaluate bone density of the proximal femur has the potential to be adversely affected by rotation along the head/condyle axis, due to the complex geometry of the region being examined. To experimentally investigate the magnitude of variation attributable to rotation, the bone density of a cadaveric femur, placed in a water bath to simulate soft tissue, was measured as a function of rotation by using a dual-energy x-ray absorptiometry (DXA) system. It was found that 10 to 15 degrees of rotation was the minimum deviation from the baseline of the femoral head perpendicular to the x-ray beam required to produce statistically significant (P = .95) changes in measured bone density. The magnitude of the variation at these rotations ranged from 2.5% for the trochanter to 5.0% for the femoral neck. Even though variation along the axis described is difficult to precisely control in the clinical setting, the magnitude of density errors attributable to these variations should not adversely affect the utility of projectional density measurement techniques.