Bone cells respond to their mechanical environment to maintain, resorb or form bone matrix;​ however, specific underlying mechanisms and mathematical principles governing mechanical adaptation remain unknown. Computational models of mechanical adaptation have relied on simplifying and often arbitrary assumptions, as the salient features of mechanical signals which control the adaptive response remain unclear A few in vivo models, such as the functionally isolated turkey ulna, allow control of the mechanical environment, providing an opportunity to systematically investigate the response of cortical bone to various loading regimens. However, in and of themselves, in vivo models cannot provide local stress/strain fields to afford testing of specifically postulated governing mechanical param eters.

Computational analyses quantified adaptive responses of the ulna at mid-diaphysis, as well as the mechanical environment throughout the ulna diaphysis during natural (homeostatic) activities and experimental regimens. The assumption of long bone symmetry was examined, establishing the fidelity of geometrical indices and contemporary imaging techniques u sed to assess adaptive changes. Finite element stress analysis demonstrated a high degree of structural symmetry. In vivo strain measurements and 24 hour observation of animal activity quantified the natural and experimental daily strain histories of the turkey ulna. hese data were used to test the daily stimulus theory of bone adaptation, which proposes a governing daily stimulus m agnitude as a function of exponentially-weighted stress m agnitudes and numbers of cycles. The daily stimulus formulation did not discriminate between strain histories which maintain and increase bone mass, suggesting the theory is oversimplified. Circumferential variation in apposition at bone formation at ulna mid-diaphysis was quantified and compared to candidate surface modeling signals formulated from three-dimensional stress analyses. Site-specific changes in peak magnitudes of 43 candidate mechanical parameters, from natural physiologic wingflap to experimental loading, were tested for linear correlation to the quantified adaptive responses on the periosteal and endosteal surfaces. Error signals which em erged as relatively strong predictors (R>0.8) of periosteal surface formation for axial loading did not predict the m aintenance response of torsional loading. These studies suggest limited utility of this deterministic approach in identifying mechanical factors initiating appositional bone formation.