For this study, bone-mineral content and bone-mineral width were measured using photon absorptiometry at eleven locations in forty excised femora and tibiae obtained from an archaeological sample. An additional seventy-nine femora were scanned at the middle of the shaft and through the femoral neck. After scanning, the bones were sectioned at each location and cross-sectional areas and other geometrical properties were determined directly from section tracings. The results indicated that locational, sex, and age-related differences in bone-mineral content were largely determined by variation in cortical area. Due to differences in bone geometry, variation in bone width does not reflect variation in cortical area, and as a consequence the use of bone width to standardize for volumetric or bone "size" differences produces misleading results in sex and age comparisons. In this study, decreases with age in the bone-mineral content and bone-mineral content:bone width ratio were similar to those observed in living populations. However, the bone-mineral content:cortical area ratio showed no significant decline with age for any cross section. Thus, the age-related changes in compact cortical bone appeared to be mainly volumetric, not densitometric.