An experiment was undertaken to obtain approximate values for the intrinsic elastic modulus of subchondral bone. Shallow spherical caps, with uniform and incrementally controlled thickness, were machined from subchondral bone in the weight‐bearing regions of 11 fresh‐frozen normal femoral head autopsy specimens. Under application of polar point loads, the measured deflections were compared with a corresponding analytical shell solution, thus allowing back‐calculation of the apparent modulus. Analogous tests were performed on similarly shaped specimens of stock Plexiglas® of known modulus in order to estimate the precision of the testing method. The aggregate results for subchondral bone showed that its intrinsic stiffness correlated inversely with nominal shell thickness, but even the thinnest (1.0 mm thick) of these shells had an apparent modulus (mean = 1.372 GN/m², SD = 414 MN/m²) well below that generally accepted for “pure” cortical bone (about 14 GN/m²). This stiffness deficit was very likely due to the presence of histologically evident marrow spaces. However, the low apparent modulus values measured in this study may not be fully representative of complex in vivo behavior, because in the testing of excised shells there is no radial compressive stress transfer to underlying cancellous bone.