The standard methodology for measuring loads in long bones is the in situ load cell, which enables direct measurements, but alters the stiffness and mass of the subject bone. Bone loading can also be calculated by applying linear beam theory to measurements from strain gauges affixed to the bone surface. The efficacy of the strain gauge method was assessed in this study by mounting three strain gauge rosettes to the midshaft of the tibia in two cadaveric above-knee leg specimens. The specimens were subjected to quasistatic axial compression tests, and then the tibia was removed and subjected to four-point bending tests. Linear beam theory for an irregularly shaped cross-section was used to calculate the axial load and bending moments in the tibia. It was possible to accurately calculate the bending moments in the bone, but the calculated axial loads appeared to be grossly in error (up to nearly 50%). This error was attributed to bone curvature and deviations from assumptions of bone homogeneity and linearity. The errors in the axial load results could be corrected by calculating an "effective" centroid for each bone, which was found to be approximately 1.5 mm away from the location of the area centroid as determined from CT scans. In spite of the error associated with calculating axial loads, this methodology shows promise for straight bones and for biomechanical experiments in which long bone bending is the parameter of greatest interest and implanting a load cell is problematic (e.g., vehicle-pedestrian tests).