On the stresses induced by flexure in a deep rectangular beam
1. When a straight cylindrical rod is bent into a circle by couples applied at its ends, the resulting state of stress is given, with sufficient accuracy for practical purposes, by the well-known theory of St. Venant. In that theory qunatities of the second and higher orders in therms of strain are neglected, and the resulting solution asserts that the stress is purely longitudinal, so that the rod may be thought of as an assembly of cylindrical fibres, each of which behaves independently of its neighbours. It is evident that this description cannot be exact; for a fibre bent into a circle cannot be kept in tension unless radial forces operate to maintain equilibrium, and in the case considered such forces can come only from actions between adjacent fibres. The apparent paradox is explained by the consideration that those action are of the second order in terms of the curvature, and accordingly are neglected in St. Venant's theory. In connection with a certain problem of elastic instability it was thought desirable to attempt a more accurate description for a particular case, namely, a rod of deep and thin rectangular section. It was found that the equations of equilibrium can be integrated independently of any simplifying assumption, and the stress-distribution determined for curvature of any magnitude. The results have no great practical importance, sice they show that St. Venant's theory gives a close approximation to the facts within that range of strains which actual materials can sustain elastically; but they have some theoretical interest, and accordingly are presented in this paper.