The purpose of this study was to determine the degree of bilateral variability in cross-sectional geometric properties of the adult canine proximal femur and to use these data to determine minimum detectable treatment effects in paired and independent experimental designs. Thirteen pairs of canine femora were sectioned at nine locations and 16 cross-sectional geometric properties were determined for each section location. The canine femur was found to be bilaterally symmetrical. For a given sample size, the magnitude of the detectable treatment effect was (a) smaller for diameters than for areas and area moments of inertia and (b) smaller within the middiaphysis than proximally. The data from this study can be used to estimate sample size requirements for experiments in which the treatment effect is determined by using the contralateral femur as a control. It was found that an increase from 3 to 7 animals would have a much greater effect on improving the sensitivity of an experiment than would an increase from 7 to 11 animals.