Several techniques are now available for quantitation of bone mass, including total body calcium by neutron activation analysis (TBC), single and dual photon absorptiometry at the radius and spine, respectively (SPA and DPA), quantitative computed tomography of the spine (QCT), and cancellous bone volume from iliac crest bone biopsies. The aims of this study were to assess the correlation among these techniques and to determine if bone mass correlated with height loss or the severity of vertebral fractures in osteoporotic women. These measurements were performed in the same group of women (n = 122) with postmenopausal osteoporosis. TBC, SPA, DPA, and QCT correlated significantly with each other (r = 0.33-0.76). The correlation between QCT and DPA improved significantly (r = 0.33-0.57; P = 0.02) when integral QCT was used instead of the usual QCT of the central vertebra. The correlation was not improved by expressing DPA in different dimensions, but was improved by including normal women. Multiple linear regression analysis showed that TBC was predicted better using SPA, DPA, and QCT than by any single variable. Cancellous bone volume correlated weakly (r = 0.10-0.26) to the other measurements. Height loss and fracture index correlated significantly with each other (r = 0.66; P less than 0.001), but the correlations with bone mass measurements were weak (r = -0.09 to -0.47). For each patient, a bone mass index was defined using z scores from the SPA, DPA, QCT, and TBC measurements; this index correlated inversely with the fracture index (r = -0.29; P = 0.009). This correlation was not significant when controlled for age. Thus, bone mass measurements correlate moderately well with each other, but poorly with severity of fracture. Other factors in addition to low bone mass may be important in determining whether bone will fracture.