This paper follows up the analysis of relativity theory begun by Margenau and Mould, by including electromagnetic theory which in their treatment was tacitly accepted. It is shown that the experiments on which Margenau and Mould rely to establish the special theory of relativity actually confirm the mutual consistency of the Maxwell-Lorentz electromagnetic theory and the special relativity theory, but throw no light on the validity of the two theories taken jointly. It is further shown that a modification of the rules of correspondence between the mathematical structure of the theories and immediate experience would bring the theories into agreement with an alternative relativity theory based on the Galilean instead of the Lorentz transformation. An experiment is suggested by which the need for such modification can be tested. A proof is then given that the rules of correspondence between the concepts of the special relativity theory (and therefore of current electromagnetic theory) and experience are not self-consistent, so that some modification of current ideas is essential. It is suggested that a generalisation of Maxwell's theory, in terms of Faraday's “ray vibrations” instead of Lorentz's static ether, might provide a satisfactory basis for a relativistic electromagnetic theory.