Altogether 426 women had their forearm bone mineral content (BMC) measured with single photon absorptiometry (SPA): one group in the early 1970s, another about 18 years later. Both groups represented purportedly healthy subjects. A third group of 328 women, measured at the same time as the second group, was from the same population but chosen by random selection (the population-based group). In the two sets of non-population-based women there was a small percentage reduction in the cortical forearm bone mass in the recent measurement as compared with the earlier measurement. This was significant only in women below 70 years of age. The forearm BMC, both cortical and trabecular, was less in all age groups of women randomly selected from the same population as the healthy control sample. This difference emphasizes the importance of selecting normative data from a population-based sample.