THE ZEEMAN EFFECT AND DISPERSION. 419 



other. When, however, the frequency of vibration of the 

 links is at all comparable with the frequency of relative 

 vibration of the parts of each double mass these parts will 

 be more or less violently agitated, and there will be a con- 

 siderable relative motion of the parts of each double mass 

 with a straining of the spring connection between them. 

 The effective inertia of the double masses in this case is 

 quite different from what it is when there is no internal 

 relative motion of their parts. The general result is that 

 when the frequency of vibration of the links is less than 

 that of the double masses a wave is propagated more and 

 more slowly the more nearly the two frequencies approach 

 coincidence. When the frequencies coincide ordinary wave 

 propagation ceases in the theoretical case, as all the energy 

 is spent in increasing indefinitely the internal vibrations of 

 double masses. In considering the case of ordinary dis- 

 persion we need not look into the matter further. Ordinary 

 dispersion consists in the propagation of waves of great 

 frequency being slower than that of waves of smaller fre- 

 quencies, and it is easy to see how the above-mentioned 

 analogy illustrates this phenomenon. We must suppose 

 the atoms of matter immersed in the ether capable of in- 

 ternal vibration at a frequency somewhat greater than that 

 of the light we are studying, and then we might expect 

 the waves of frequencies approaching those of the atoms to 

 be propagated more slowly than waves of slower frequencies 

 which do not agree so nearly with that of the atoms. Two 

 things will determine the amount by which the matter will 

 alter the velocity of propagation of the light. One will be 

 the amount of interconnection between the matter and the 

 ether. We might naturally expect some kinds of matter to 

 be only very slightly acted on by the vibrations of the 

 ether in which it is immersed, while other kinds of matter 

 would be much acted on by these vibrations. The reaction 

 of the matter on the ether and on the rate of light propa- 

 gation would be in the first case small, and in the second 

 case large : the first kind of matter would generally be more 

 highly refractive and dispersive than the second. The 

 other thing that will determine the amount of dispersion 



