The capacity of dual x-ray absorptiometry and quantitative ultrasound to discriminate bone loss and to predict the mechanical and microarchitectural properties of cancellous bone in an animal model of osteopenia was evaluated. Thirty-five female Sprague-Dawley rats (10 months old) were randomized into three groups: baseline group, 10 rats killed at the beginning of the study; ovx group, 15 rats ovariectomized; and sham group, 10 rats sham operated. At the beginning and end of the study, all the animals underwent osteosonography to record the proximal tail (C3 vertebra) bone speed of sound. Sixteen weeks after surgery, the animals were euthanized and the L5-6 lumbar vertebrae of each rat were excised for densitometric, biomechanical (compression test), and histomorphometric studies. Significant differences were found among the groups for final speed of sound (p = 0.01). The L5 bone mineral density of the ovx group decreased by 12.1% (p = 0.049) and 12.6% (p = 0.035) compared, respectively, with baseline and sham groups. The biomechanical parameters of the ovx group decreased by 15-47% compared with the other groups, showing significant differences between the ovx and sham groups both for maximal stress (p = 0.026) and elastic modulus (p = 0.013). Histomorphometric parameters of the ovx group showed significant decreases in comparison with other groups. Logistic regression analysis showed that dual X-ray absorptiometry and quantitative ultrasound discriminate ovariectomized and healthy rats with a similar capacity, classifying correctly all rats used in the model in a range of 61-70%. This similar capacity seems to derive from two different capacities to detect bone changes. Dual X-ray absorptiometry, depending on bone mineralization and density, is able to detect modifications in bone stiffness and strength, confirmed also by the correlation with biomechanical data. On the contrary, quantitative ultrasound seems to depend more on cancellous bone microarchitecural changes because it is correlated to histomorphometric parameters.