Juvenile BALB/c mice were used as a model system to test the effects of various loading and exercise regimens on the growth and development of femora. Six treatments and three controls were used to document changes in geometric, mechanical, and material properties of the femora associated with strength. In each age-matched experiment, body weight and the strength, length, anterior and posterior diameters, cross-sectional area, moments of inertia in the anteroposterior and lateromedial directions, cortical wall thickness, and mineral content of the femora were assessed and found to vary significantly among treatment groups. An adaptive interpretation of these data was provided by calculating Pearson correlation coefficients between moment at failure (one measure of strength) and each geometric, mechanical and material property of the femora that contributes to strength. We make the assumption that at the termination of the experiment the greater the coordination between changes in strength and changes in the parameters that contribute to strength (the greater the number of correlations), the more adaptively modeled the femora are. Adaptive modeling here refers to the manner in which the femora grow and develop (adapt) under a given treatment regimen. Absolute strength of whole femora was reflected by our measure of adaptive modeling in all groups with one exception. In each experiment, the voluntary exercise controls were the most adaptively modeled. The least adaptively modeled groups also showed a general retardation of growth. It appears that juvenile mouse femora demonstrate a wide range of responses to different conditions of loading and exercise and that some of these changes are likely permanent. Moreover, at least two major variables--1) mechanical loading and 2) glucocorticoid mediated psychological stress--appear to contribute to the differences seen between the treatment groups.