4 i8 SCIENCE PROGRESS. 



than any that have been satisfactorily measured ; perhaps 

 some phenomena may be observed in reference to Bec- 

 querell or X rays depending on their wave lengths being 

 comparable to the distances between molecules in solids. 

 Hence it is not in the direction of this theory that we 

 should expect to find an explanation of the Faraday effect. 

 The theory of dispersion that seems to explain very 

 satisfactorily the phenomena is one that attributes it to a 

 comparability between the frequency of vibration of the 

 liofht and of the atoms of matter : to a resonance of the 

 matter to the vibrations of the light passing through it. 

 It is not very easy to put the dynamical basis of this theory 

 in general terms in a simple manner. The phenomena of 

 forced vibrations are not familiar to most persons, and it 

 would take a long explanation to bring the matter clearly 

 before those not already familiar with these phenomena. 

 The part of the phenomena on which ordinary dispersion 

 depends might be illustrated by the following experiment. 

 Suppose a long chain stretched between two points. We 

 are all familiar with the way in which a wave can be pro- 

 pagated along such a chain. The rate of propagation is 

 independent of the length of the wave so long as it is not 

 comparable in length with the links of the chain. If the 

 links have masses of lead attached to them the result will 

 be that these long waves will all be propagated more slowly 

 than before, but all the same amount more slowly. This 

 may be taken as analogous to the slower propagation of 

 light in transparent matter than in free ether. It will com- 

 plicate this simple result if the leaden masses, instead of 

 being simple masses each firmly attached to a link, are 

 each made up of two masses fastened together by a spring 

 so as to be capable of vibrating independently of the links. 

 The effective inertia of these double masses will be the 

 same as if they were rigidly fastened together so long as 

 the vibrations of the links are very slow compared with the 

 independent vibrations of the double masses due to their 

 parts being connected by springs. When the links vibrate 

 slowly the double masses will move together, and there will 

 be very little motion of one of their parts relatively to the 



