It is shown that the conventional correspondence procedure used in the transition from Newtonian to the Einsteinian theory of gravitation is not unique. A new procedure of correspondence whereby the theory is required to reduce to a special relativistic limit before it reduces to the Newtonian limit is described. It is shown that the special relativistic correspondence leads to first and second order requirements, whereas the Newtonian correspondence leads only to first order requirements. The resulting theory is logically plausible and experimentally viable.