To investigate associations among methods for noninvasive measurement of skeletal bone mass, we studied 40 healthy early postmenopausal women and 68 older postmenopausal women with osteoporosis. Methods included single- and dual-energy quantitative computed tomography (QCT) and dual-photon absorptiometry (DPA) of the lumbar spine, single-photon absorptiometry (SPA) of the distal third of the radius, and combined cortical thickness (CCT) of the second metacarpal shaft. Lateral thoracolumbar radiography was performed, and a spinal fracture index was calculated. There was good correlation between QCT and DPA methods in early postmenopausal women and modest correlation in postmenopausal osteoporotic women. Correlations between spinal measurements (QCT or DPA) and appendicular cortical measurements (SPA or CCT) were modest in healthy women and poor in osteoporotic women. Measurements resulting from one method are not predictive of those by another method for the individual patient. The strongest correlation with severity of vertebral fracture is provided by QCT; the weakest, by SPA. There was a high correlation between single- and dual-energy QCT results, indicating that errors due to vertebral fat are not substantial in these postmenopausal women. Single-energy QCT may be adequate and perhaps preferable for assessing postmenopausal women. The measurement of spinal trabecular bone density by QCT discriminates between osteoporotic women and younger healthy women with more sensitivity than measurements of spinal integral bone by DPA or of appendicular cortical bone by SPA or CCT.