We evaluated age- and sex-specific differences in bone density at a variety of skeletal sites in a population-based sample of 348 men (age 22-90 years) and 351 women (age 21-93 years) by dual-energy X-ray absorptiometry. Several patterns of age-related bone loss were observed, but adjustments for height (or, where possible, calculation of bone mineral apparent density) dampened the apparent rate of bone loss from most skeletal sites as judged from cross-sectional measurements at baseline. Cross-sectional data also overestimated the rate of bone loss observed longitudinally over 4 years at many sites, particularly the hip and spine; conversely, in some forearm regions, cross-sectional rates of loss underestimated the bone loss seen prospectively. Longitudinal rates of bone loss were generally greater among individuals age 70 years or older compared with younger men and women, but some of the latter effect was due to a greater proportion of younger women on hormone replacement therapy, whose rate of bone loss was generally less. These observations highlight the limitations of cross-sectional data for defining the patterns of bone loss over life at different skeletal sites.