THE ZEEMAN EFFECT AND DISPERSION. 417 



magnetised matter on the light. Magnetic force changes 

 matter in some way owing to which the magnetised matter 

 transmits right- and left-handed circularly polarised beams 

 at different rates along the direction of magnetisation. 

 Without matter there is no known action of magnetic force 

 on light. Another thing to be observed is that the action 

 is negligible for very long ether waves and only becomes 

 of importance for the short waves which are in the neighbour- 

 hood of the visible part of the spectrum. The phenomenon 

 is much more closely allied to dispersion than to refraction. 

 The longest ether waves are not propagated at the same 

 rate in matter as in free ether, they are largely refracted. 

 Long ether waves are, however, all propagated at nearly 

 the same rate. The differences of rate of propagation that 

 produce the phenomena of ordinary dispersion are only 

 of importance in the case of the short waves which are in the 

 neighbourhood of the visible part of the spectrum. In the 

 case of both the effect Faraday observed, and of dispersion, 

 the action is roughly inversely proportional to the square of 

 the wave length, thus leading one naturally to expect that 

 they may be allied phenomena. 



If, with this clue, we were to seek for some explanation 

 of the Faraday effect we should study the various theories 

 that have been propounded to explain dispersion. The 

 oldest of these explains dispersion by the hypothesis that 

 the distances between the molecules of matter are compar- 

 able with the lengths of the waves of light. In the case of 

 wave lengths, for which this is true, there must exist dis- 

 persion phenomena, and until some rough estimate had 

 been made of the distances between molecules of matter 

 there seemed good reason to suppose that the phenomena 

 of ordinary dispersion were due to this cause. The theory 

 explained the laws of ordinary dispersion as well as could 

 be expected from a rough theory, which took no account 

 of the structure of the molecules themselves. Now, how- 

 ever, it is known that the distances between the molecules 

 in transparent solids is a great deal too small to explain 

 ordinary dispersion. Dispersion due to this cause must be 

 concerned with waves of very much shorter wave length 



