alter the free periods of the spectral lines of the substance. Bat the actual phenomena do not seem to be thus reciprocal. On the electric theory of light it is only the dispersion in material media that arises from direct influence of the free molecular periods, the main refraction arises from the static dielectric coefficient of the material, which is not connected with the periods of molecules.* From the phenomena of magneto-optic reflexion it may be shown that, on the hypothesis that the Faraday effect is due to regular accumulated influences of the individual molecules, it must be involved in the relation between the electric force (PQK) and the electric polarisation of the material (f'g'li), of type / ' K - l p 4 tt
where (ciC2c3) is proportional to the impressed magnetic field. This relation, interpreted in the view that the electric character of a molecule is determined by the orbits of its electrons, simply means that the capacity of electric polarisation of the molecule depends on its orientation with regard to the imposed magnetic field, that, in fact, the static value of K, depending on the molecular configurations just as much as do the free periods, is altered by the magnetic field. This relation agrees with the main feature of rotatory dispersion, namely, that it roughly follows the law of the inverse square of the wavelength. The specific influence of the molecular free periods, that is, of the ordinary dispersion of the material, on the Faraday effect, is presumably a secondary one; though it, too, follows the same law for different wave-lengths, in the case of substances for which Cauchy’s dispersion formula holds good. It is this latter part of the Faraday effect that is reciprocal to Dr. Zeeman’s phenomenon.
The question is fundamental how far we can proceed in physical theory on the basis that the material molecule is made up of revolving electrons and of nothing else. Certain negative optical experiments of Michelson almost require this view ; at any rate, they have not been otherwise explained. It may be shown after the manner of ‘Phil. Trans.,’ 1894, A, p. 813 (and Dr. Zeeman’s calculation, in fact, forms a sufficient indication of the order of magnitude of the result), that in an ideal simple molecule consisting of one positive and one negative electron revolving round each other, the inertia of the molecule would have to be considerably less than the chemical masses of ordinary molecules, in order to lead to an influence on the period, of the order observed by Dr. Zeeman. Bat then a line in the spectrum may be expected to arise rather from one of the numerous epicycles superposed on the main orbits of the various electrons in the molecule than from a main orbit itself.
- Loc. cit., 1 Phil. Trans.,’ 1894, A, p. 820; and 1895, A
