THE ZEEMAN EFFECT AND DISPERSION. 427 



whose frequencies are one greater and the other less than 

 that of the unchanged vibration. The radiations emitted 

 by these circular components in the plane of their motion 

 are plane polarised in a plane at right angles to the un- 

 changed one. The effect observed in the spectroscope is 

 the production of two lines, one on each side of the un- 

 changed one, and both plane polarised in a plane at right 

 angles to that of this unchanged one. Further, on consider- 

 ing the direction of rotation of the circularly polarised 

 beams emitted along the lines of magnetic force, it can be 

 shown that the beams must be principally due to the rota- 

 tion of a negative electron. 



All this direct observation of the effect of magnetic force 

 on the vibrations of molecules confirms the theory already 

 expounded as to what ought to take place in order that we 

 might explain the Faraday effect by a resonant theory of 

 dispersion. It proves that there really exists the action of 

 magnetism on the vibrations of the molecules that that 

 theory postulated, and consequently proves that here is a 

 vera causa for a Faraday effect. All that is further required 

 is to show that it is a sufficient cause for the observed Fara- 

 day effect. Unfortunately the data required for this con- 

 firmation of the theory are not available. We should know, 

 in the case of at least one gas, the amount of the Faraday 

 effect, of the Zeeman effect, and in addition the law of dis- 

 persion by the gas sufficiently accurately to calculate the 

 frequency of the principal absorption that controls its dis- 

 persion. The Zeeman effect and Faraday effect have not 

 been measured in the same gas. Besides this, the Zeeman 

 effect is not the same for the various lines in a single spec- 

 trum, and we know no law for calculating its effects on one 

 line from observations on another. What we require in our 

 calculation is the Zeeman effect on the absorption band 

 that controls dispersion, and this has not been observed in 

 any gas, the band being far up in the ultra-violet. For these 

 reasons there is little prospect of any except a very rough 

 comparison between theory and experiment being possible 

 for some time. When we make a roucrh estimate of the 



possible Faraday effect from the observed amounts of the 



29 



