Extending the paper of Slater (Phys. Rev. 34, 1293, 1929) on complex spectra, it is pointed out that assignment of definite electron configurations to spectral terms is an approximate procedure and only has meaning when the multiplet systems of the several configurations are widely separated. The effect of including spin terms is sketched. Non-diagonal matrix elements for the $N$-electron problem are reduced to corresponding elements for the 2-electron problem, as Slater did for the diagonal elements. Two-electron jumps occur because of the fact that spectral terms may not be precisely labelled by means of electron configurations.

• Received 27 August 1930

DOI:https://doi.org/10.1103/PhysRev.36.1121