On the Dispersion of Light in Refracting Media. 333 



efficients of direct friction nmst be introduced into the calculus. 

 In fact the success of this speculation depends absolutely on 

 the presence of the terms conditioned by friction ; and if these 

 be eliminated by putting the friction-coefficient =0, the dis- 

 persion-curve obtained loses all resemblance to that required 

 by experiment. 



Helmholtz's treatise, however, appears to me an explanation 

 ad hoc, inasmuch as it neglects the two different kinds of 

 elliptic polarization established by experiment (of glass and 

 metals, the positive and the negative), and also the modifica- 

 tion effected in the velocity of propagation by the translation 

 of a medium, or gives no account of them, and, lastly, as it 

 contradicts the theories of reflection which have been hitherto 

 held by Fresnel, Cauchy, and myself, all of which require, at 

 least for the boundary surface, a simple relation between the 

 vires viva? and the ratio of refraction. 



To me it becomes difficult, especially after the results of my 

 last work, to abandon the representation that a refraction-index 

 variable with the angle of incidence belongs to the metals, and 

 that absorption plays, instead of a primary, only a secondary 

 part. After much pains I have at last been compelled to de- 

 cide that it is impossible to arrive at expressions serviceable in 

 every direction, even by extending Helmholtz's assumptions by 

 the addition of new terms ; while these hypotheses appear to 

 me in part not unobjectionable, even theoretically. 



For experiment shows the course of the dispersion-curve to 

 be independent of the condition of the aggregate. If now we 

 consider that only towards very rapid motions does the aether 

 behave as a solid body, while its resistance to slow displace- 

 ments, even those of the light particles of gas flying about in 

 space, entirely vanishes, and that, in agreement with this, 

 according to observations of aberration, the translation of a 

 medium which takes place perpendicular to the direction of 

 the rays leaves the velocity of propagation of them unaffected, 

 the mechanism of the reciprocal action between the aethereal 

 and corporeal particles may be in reality more complicated than 

 Helmholtz assumes it to be. According to our simple consi- 

 deration the influence of the corporeal particles upon large 

 oscillation-periods becomes just =0, and hence, for all aggre- 

 gate forms, the corresponding velocity of propagation =1, 

 while Helmholtz's theory gives this velocity, at least for gases, 

 as 0. 



If now I proceed to set up a theory of my own, in my 

 opinion the difficulties indicated may be avoided in the follow- 

 ing simple way : — 



2. We confine ourselves in the following to media at rest, 



