Estimation of torsional rigidity in primate long bones

Comparative studies of long bone biomechanics in primates frequently use the polar moment of inertia (J ) as a variable reflecting overall mechanical rigidity, average bending rigidity, or resistance to torsional shear stresses. While the use of this variable for characterizing the first two properties is appropriate, it is potentially a highly misleading measure of torsional resistance. Errors result from violations of assumptions required for the use of the polar moment of inertia; in particular, the predictive utility of J diminishes with departures from axial symmetry (i.e., a cylindrical cross-sectional shape). The magnitude of these errors is estimated both theoretically and experimentally. It is argued that the use of the polar moment of inertia for estimating long bone torsional rigidity should be restricted to samples of relatively invariant and/or cylindrical geometry. Alternative measures for torsional resistance are evaluated and reviewed.