An exponential increase in hip fracture risk is an apparently immutable consequence of aging in our society. Approximately 1.66 million hip fractures occurred worldwide in 1990, and this number is expected to increase markedly as the elderly population grows. From an engineering viewpoint, a hip fracture is a structural failure of the proximal femur resulting when loads applied to the bone exceed its structural capacity. The objective of this thesis was to improve current understanding of hip fracture etiology by investigating factors contributing to the structural capacity of the femur. This goal was accomplished using two research strategies:​ material level studies examining multiaxial failure properties of trabecular bone, and finite element studies at the whole bone level aimed at elucidating the relative contributions of age-related bone loss and fall mechanics to hip fracture risk. Using an analytical approach applicable to cellular materials, failure surfaces for general, three-dimensional states of stress were derived for trabecular bone of any density and degree of transverse isotropy. Comparison with surfaces derived experimentally demonstrated that the microstructural failure mechanisms used in the model do contribute to multiaxial failure behavior at the continuum level. For studies at the structural level, an anatomically accurate finite element model of the proximal femur was generated using geometric and densitometric information from Quantitative Computed Tomography (QCT). The structural capacity of the femur was predicted under load vectors representing in vivo falls impacting on the hip, revealing that variations in impact direction can dramatically influence load-bearing capacity. Using a novel algorithm allowing direct comparison of QCT data from femurs with different geometry, three-dimensional patterns of bone loss associated with osteopenia and osteoporosis were determined. The structural consequences of bone loss were assessed by applying these highly non-uniform density reductions to the finite element model. The results suggest that clinically "uniform" bone loss (i.e. that which appears uniform based on two-dimensional clinical bone density exams) causes progressively larger percent reductions in structural capacity, and that inter-region differences in the rate of bone loss can have a dramatic effect on structural capacity.