Finite-element modeling provides a full-field method for describing the stress environment of the skull. The utility of finite-element models, however, remains uncertain given our ignorance of whether such models validly portray states of stress and strain. For example, the effects of boundary conditions that are chosen to represent the mechanical environment in vivo are largely unknown. We conducted an in vitro strain gauge experiment on a fresh, fully dentate adult mandible of Macaca fascicularis to model a simplified loading regime by finite-element analysis for purposes of model validation. Under various conditions of material and structural complexity, we constructed dentate and edentulous models to measure the effects of changing boundary conditions (force orientation and nodal constraints) on strain values predicted at the gauge location. Our results offer a prospective assessment of the difficulties encountered when attempting to validate finite-element models from in vivo strain data. Small errors in the direction of load application produce significant changes in predicted strains. An isotropic model, although convenient, shows poor agreement with experimental strains, while a heterogeneous orthotropic model predicts strains that are more congruent with these data. Most significantly, we find that an edentulous model performs better than a dentate one in recreating the experimental strains. While this result is undoubtedly tied to our failure to model the periodontal ligament, we interpret the finding to mean that in the absence of occlusal loads, teeth within alveoli do not contribute significantly to the structural stiffness of the mandible. © 2005 Wiley-Liss, Inc.