It is unclear what proportion of the variance in bone density in elderly males is accounted for by testosterone status. We studied 112 ambulatory, elderly volunteers (mean age 71.7 years) and determined free testosterone (FT), as well as bone density measurements by photon absorptiometry at multiple sites. Our studies of 35 of these subjects 4 years later included morning FT and dual energy X-ray absorptiometry. There were no significant correlations between FT and bone density at multiple scanning sites with the effects of age partialed out. We suspect that our inability to detect a significant effect of FT on bone density was related to the relative strength of other determinants of bone density, as well as to the fact that FT values are far more dynamic than bone density.