Alternatively, one can work in the interaction representation in the Heisenberg-Pauli formalism. In the calculation of the S-matrix this differs from the σ-formalism only in the form of the terms which can safely be ignored (reference 3).

These rules are essentially the same as those given by Feynman

The author is grateful to Drs. D. Feldman and F. Rohrlich for helpful correspondence on rule (vi).

This result was given non-relativistically by

and has been obtained relativistically by a different method by F. Rohrlich (to be published). The author is indebted to Professor Pauli for informing him of Dr. Rohrlich's result.