A mathematical procedure is described for transforming optical functions from one set of variables to another, irrespective of the order to which the functions are developed. The procedure is explained, and several problems to which it can be applied are outlined.