Osteocytes are believed to be the mechanical sensor cells in bone. One potential physical mechanism for the mechanosensing process is that osteocytes directly sense the deformation of the substrate to which they are attached. However, there is a fundamental paradox in this theory: tissue-level strains in whole bone are typically <0.2%, yet an extensive range of in vitro experiments show that dynamic substrate strains must be at least an order of magnitude larger in order for intracellular biochemical responses to occur. Recently, a theoretical model was developed (You et al. J. Biomech., 2001; 34:1375–1386) that provides a possible mechanism by which mechanical loading-induced fluid flow in the lacuno-canalicular system, under routine physical activity, can produce cellular-level strains on the osteocyte processes that are at least one order of magnitude larger than bone tissue deformations. This would resolve the fundamental paradox mentioned above. In this work we experimentally confirm and quantify the essential ultrastructural elements in this model: 1) the presence of the transverse elements that bridge the pericellular space surrounding the osteocyte process, which interact with the fluid flow and lead to an outward hoop tension on the process; and 2) the presence of bundled F-actin in the osteocyte processes, which resists the outward hoop tension and limits the cell process membrane deformation. Morphological data to support these assumptions are scant. Special staining techniques employing ruthenium III hexamine trichloride (RHT) were developed to elucidate these structures in the humeri of adult mice. Anat Rec Part A 278A:505–513, 2004. © 2004 Wiley-Liss, Inc.