New technology for noninvasive measurement of bone mass has enabled many studies of bone mass and its relationship to fracture, which challenge the view that bone mass is the only relevant factor in the etiology of fractures. Several studies have reported ROC curves that generally show values of about 80%. No convincing evidence suggests that one technique is superior to another. The reported relative risks or odds ratios for a fracture usually range between 1.2 and 2.5 per SD. There is no doubt that the risk of a fracture increases as the bone density decreases. However, even with a low bone mass, the risk of not fracturing a bone over the next year is over 90%. Most of the data suggest that patients with severe vertebral fractures have lower bone mass than those with mild fractures, but some women with similarly low bone mass have mild or no fractures. The weight of the evidence suggests that age has an effect on fracture incidence which is independent of bone mass. Trauma is such a major factor that it is surprising to find almost no studies that have controlled for it. The relationship between bone mass and bone failure is strong, but other factors must also be contributing to the bone failure which, like heart failure or renal failure, is a complex, multifactorial disease.