where variables subscripted with e or p denoteelectron and nuclear variables, and Z is the nuclearcharge.In working withthis Hamiltonian we must keep inperspective its approximate nature. Consistencyrequires that in applications the Breit potentialshould only be treated in first order, since secondordereffects would give results of the same orderas terms which have been neglected in Eq. (l).Keeping this limitation in mind, we may derivefrom Eq. (l) a Hamiltonian H which acts only inthe upper component space of the nucleus. In effectwe are treating the nucleus as nonrelativistic.This approximation is valid in most hydrogenicatoms (positronium excluded) because of the smallnessof the ratio m/M.Dropping nuclear-spin-dependent terms, whichmay be handled by perturbation theory, we obtainto first order in M 'H = n, ~ p, + P,m +