378 ON LIGHT. 



mond, and all those bodies which possess what is called 

 the adamantive lustre (a consequence of such high refrac- 

 tive power) it is considerable. From such bodies accord- 

 ingly it is not possible, at any angle of incidence to obtain 

 a reflected ray completely polarized in one plane. And 

 when we come to reflexion from polished metals,* the 

 ellipticity becomes very considerable. In consequence, 

 only a very imperfect polarization of the reflected light 

 in the plane of incidence can be obtained by reflexion 

 from any metalic surface at any angle. 



(156.) In all the above enumerated cases, the degree 

 of ellipticity increases with the reflective power of the 

 medium of which the reflecting surface is constituted ; 

 which itself stands in intimate connexion with the mag- 

 nitude of the refractive index. It might naturally, there- 

 fore, be expected to attain its maximum possible amount, 

 or that the ellipse should become a circle in the case 

 of total reflexion. This can only take place, however, 

 when the reflexion is made on the internal surface of a 

 transparent medium. This accordingly happens in the 

 case of a beautiful experiment of M. Fresnel, who found 

 that a parallelepiped of glass, t A B c D, fig. 15, being 

 cut and polished, having the acute angles at A and D, 



* All metals, even the densest, are in some slight degree trans- 

 parent, and all have enormously large refractive indices. The 

 transparency of gold is perceptible in gold leaf, which transmits a 

 green light. That of silver is perceptible in the thin films deposited 

 on glass in Liebig's process for silvei'ing mirrors the transmitted 

 light being bluish. 



f The glass used was that known in France as " Verre de St. 

 Gobain." 



