On the choice of the electron-electron potential in relativistic atomic physics

In the calculation of relativistic effects in the structure of many-electron atoms there are two popular choice for the operator V used to represent the electron-electron interaction. One of these, V(I), is associated with the use of the Coulomb gauge propagator for photons; the other, V(II), is associated with the use of the Feynman gauge propagator. In contrast to V(I), the nonperturbative use of V(II) leads to energy levels which are already incorrect in order (alpha exp 4)m. This fact provides a quantitative argument for preferring V(I) to V(II).