Electric Conduction in Metals

THIS note has reference to the very interesting paper by Mr. R. H. Fowler in the Supplement to NATURE of Oct. 18 last on the conduction of electricity in metals. In this he refers to the difficulty in connexion with Drude's model in that the free electrons in the spaces should behave like a perfect gas, and so affect the specific heat relations As experience shows that this is not so, the inference is drawn that the dynamics is wrong and we must invent a new one. But another inference would seem more natural: that the conclusion that the electrons behave like a perfect gas is wrong and we must seek the reason why. I think this is not far to seek, and the error lies in supposing that the same electrons are continuously in the free space. This may be illustrated by taking the theory of emission as proposed recently by me in the Philosophical Magazine (8, 108; 1919). Shortly put—the nucleus decides what particular orbits of electrons round it are to fall in, the energy of fall goes to the nucleus, and the nucleus radiates according to the nh law. Conversely, the reverse process also takes place. When radiations fall on the atom it absorbs those of its special periods, and when it has absorbed the quota hσ it raises an electron to a corresponding outer orbit. This then falls in and its energy is radiated again by the nucleus. Amongst these operations, some will be due to free electrons falling in, and a similar set being set free. Consequently, the free electrons in the interatomic spaces will be continually passing in and out, and they never have time to take up the energy due to the temperature of the space. The time required for this must be considerable, measured in terms of the mean interval between collisions.