In clinical practice, decisions must be made about whether and how to convert to newer technologies. To address this issue, two separate studies were conducted. We evaluated the relationships between results of lumbar spine measurements using two dual photon absorptiometry (DPA1 and DPA2) instruments and one dual energy X-ray (DXA) instrument with the same subjects (49 volunteers), and also in 65 patients who were measured on the DPA1 and DXA machines. Second, we measured the lumbar spine and the proximal femur in three groups of 12 female volunteers three times on one instrument within 1 week. We purposely simulated a busy clinic setting with different technologists, older radioactive sources, and a heterogeneous patient group. The comparison study indicated a significant difference between the mean bone density values reported by the machines, but the results were highly correlated (R2 = 0.89-0.96). The short-term precision errors (coefficients of variation) differed among the instruments, ranging from 1.3% (DXA of the spine) to 5.1% (DPA1 of the spine), and in the femoral neck, 2.3% and 2.4% (DXA and DPA1, respectively) versus 3.5% by DPA2. This study emphasizes the differences between instruments, the potential for greater error in busy clinic environments, and the apparent superiority of dual energy X-ray absorptiometry under these less than ideal conditions.