Osteoporosis is currently defined in terms of low bone mass. However, the source of fragility leading to fracture has not been adequately described. In particular, the contributions of bone tissue properties and architecture to the risk or incidence of fracture are poorly understood. In an earlier experimental study, it was found that the architectural anisotropy of cancellous bone from the femoral heads of fracture patients was significantly increased compared with age- and density-matched control material (Ciarelli et al., J Bone Miner Res 15:​32–40;​ 2000). Using a combination of compression testing and micro-finite element analysis on a subset of cancellous bone specimens from that study, we calculated the hard tissue mechanical properties and the apparent (macroscopic) mechanical properties. The tissue modulus was 10.0 GPa (SD 2.2) for the control group and 10.8 GPa (SD 3.3) for the fracture group (not significant). There were no differences in either the apparent yield strains, percentages of highly strained tissue, or the relationship between apparent yield stress and apparent elastic modulus. Hence, a difference in the tissue yield properties is unlikely. At the apparent level, the fracture group had a significantly decreased transverse stiffness, resulting in increased mechanical anisotropy. These changes suggest that bone in the fracture group was “overadapted” to the primary load axis, at the cost of fragility in the transverse direction. We conclude that individuals with a history of osteoporotic fractures do not have weaker bone tissue. Architectural and mechanical anisotropy alone renders their bone weaker in the nonprimary loading direction.