Abstract
(i) Gauge invariance in the presence of nonlocal potentials is reviewed. It is shown that the uncoupled Hartree‐Fock method is not gauge invariant. (ii) Gauge invariance of variational calculations is reviewed and it is shown that the coupled Hartree‐Fock approximation is gauge invariant. (iii) A procedure for choosing a best gauge is suggested and in a special case is shown to reduce to a method recently proposed by Karplus and Kolker.