Key points are not available for this paper at this time.
By using metallic energy levels extrapolated from copper to nickel, the energy difference between a nonmagnetic and a ferromagnetic state with permanent magnetic moment is calculated for nickel, and it is shown that the ferromagnetic state is the stable one. Both saturation magnetic moment and Curie point are calculated, in agreement with experiment within the limits of error of the calculation. Extrapolation further into the iron group, though less justified than to nickel, indicates that ferromagnetism should persist in that group down approximately to iron. The criterion for ferromagnetism previously suggested by the author, the existence of inner unfilled electron shells (the 3d), small in proportion to their distance apart, is justified. The calculation is not made according to Heisenberg's method, which is considered to be unsuitable for application to ferromagnetism, except in its general principle of explaining the energy of orientation of elementary magnets in terms of exchange energy.
J. C. Slater (Wed,) studied this question.