Keyak and Skinner¹ state that the adequacy of a finite element mesh may be verified by confirming that convergence of strain energy has been achieved. The paper concludes by stating that convergence of strain energy does not ensure that a particular mesh is adequate for producing accurate stress/strain results, and that only qualitative results can be achieved from 3D models of bone. This conclusion is challenged and the effect of element size on stress convergence is clarified. The influence of element size on stress convergence in regions of surface irregularity is studied and current methods are discussed that facilitate convergence.