226 FRESNEL. 



that a ray may have the same properties round all points 

 of its circumference, and yet not be common light. To 

 show by a single example that these two species of light 

 comport themselves differently, and ought not to be con- 

 founded, I Avill observe that, in undergoing double re- 

 fraction, a natural ray after traversing a plate of crystal 

 gives two white images, wliile under the same conditions 

 the ray of Fresnel is decomposed into two beams, each 

 hrilliantly coloured. 



This new modification, which, having no reference to 

 the different sides of a ray, has been designated circidar 

 polay-ization, can be impressed upon rays ordinarily 

 polarized, by making them undergo two successive total 

 reflexions from the internal surfaces of a piece of glass 

 suitably formed.* The pleasure of having his name as- 

 sociated with a new kind of polarization hitherto unsus- 

 pected, would probably have sufficed for the vanity of 

 an ordinary experimenter, and his researches would not 

 have extended beyond that point. But Fresnel was 

 actuated by more elevated sentiments ; in his eyes 

 nothing seemed to have been done while any thing re- 

 mained to do. He sought, therefore, if there were not 



the double refracting crystal, the two images in the former case will 

 be always white, in the later coloured. 



* In the instance mentioned, Fresnel showed, by a remarkable in- 

 stance of theoretical prediction, that a ray polarized at 45° to the 

 plane of incidence, and twice reflected internally from glass, will 

 emerge in the condition of two ra\-s polarized in planes at right 

 angles, and one retarded by one fourth of a wave-length behind the 

 other; these being superimposed will, by mathematical consequence, 

 give rise to vibrations, no longer plane, but performed in circles; or in 

 ellipses, if the retardation be any other fraction of a wave-length. 

 Such a piece of glass is called Fresnel's Rhomb. The course of the 

 ray will be apparent by inspection of the annexed diagram, which 

 needs no further explanation. 



