Understanding of the pathogenesis of bone fragility in men requires knowledge of its structural basis. There is no evidence that gender differences in fracture rates are explained by gender differences in bone mineral content (BMC) or areal bone mineral density (BMD). This is an untested assumption. The BMD measurement integrates the modeling and remodeling that occurs on the periosteal and endosteal surfaces of bone during growth and aging. The size, shape, and architecture of the bone so formed determine its breaking strength. None of these three-dimensional structural components is "seen" by the dual photons of the densitometer. Men and women attain a similar peak vertebral height during growth. Vertebral width is greater in men, conferring higher BMC and areal BMD, but trabecular number and thickness (trabecular volumetric BMD) is no greater in men than women. Blacks have shorter vertebra than whites, and vertebral width is similar. Trabecular thickness is greater in blacks than whites. Thus, at peak, gender differences in vertebral strength are likely to be size, not BMD, dependent. Racial differences in vertebral strength are likely to be BMD, not size, dependent. Greater periosteal expansion during growth in males than females, and blacks than whites, establishes the gender and racial differences in peak bone size. Men have wider long bones than women. Blacks have wider long bones than whites. The proximity of the endocortical surface to the periosteal surface determines peak cortical width, which is similar in men, women, blacks, and whites. It is the greater distance of the cortical mineral mass from the neutral axis of a long bone in males than in females, in blacks than in whites, and in men with, than men without, fractures, that partly accounts for the greater bone strength in the first mentioned in each group. Thus, at peak, racial and gender differences in long bone strength are likely to be size, not BMD dependent. Trabecular bone loss is similar in men and women. Loss of connectivity is greater in women. Endocortical resorption is greater in women than men, but men lose less cortical width because subperiosteal apposition during aging is greater in men than in women offsetting endocortical resorption. Men with spine fractures have smaller vertebrae because vertebral width is less. Men with hip fractures have smaller femoral neck width. In both types of fractures, there is less bone in the smaller bone-reduced volumetric BMD. The relative contributions of reduced accrual during growth, excessive bone loss during aging, or both to the deficit in volumetric BMD are undefined. No antifracture efficacy trials have been done in men. Reasonable approaches to treatment include the use of testosterone in hypogonadal men, and vitamin D if vitamin D deficiency is present. Calcium supplements may slow endocortical bone loss. Bisphosphonates may increase BMD.