The premise that bones grow and remodel throughout life to adapt to their mechanical enironment is often called Wolff's law. Wolff's law, however, is not always true, and in fact comprises a variety of different processes that are best considered separately. Here we review the molecular and physiological mechanisms by which bone senses, transduces, and responds to mechanical loads, and the effects of aging processes on the relationship (if any) between cortical bone form and mechanical function. Experimental and comparative evidence suggests that cortical bone is primarily responsive to strain prior to sexual maturity, both in terms of the rate of new bone growth (modeling) as well as rates of turnover (Haversian remodeling). Rates of modeling and Haversian remodeling, however, vary greatly at different skeletal sites. In addition, there is no simple relationship between the orientation of loads in long bone diaphyses and their cross-sectional geometry. In combination, these data caution against assuming without testing adaptationist views about form-function relationships in order to infer adult activity patterns from skeletal features such as cross-sectional geometry, cortical bones density, and musculo-skeletal stress markers. Efforts to infer function from shape in the human skeleton should be based on biomechanical and developmental models that are experimentally tested and validated.