224 FRESNEL. 



fests itself in a different way from that of difference of 

 intensity. The rays subjected to it, for example, always 

 give two images in traversing calc spar ; but these images 

 are each entirely tinted with a bright and uniform colour. 

 Thus, though the incident light may be white, the ordi- 

 nary ray may be entirely red, orange, yellow, green, blue, 

 or violet, according to the direction in which the prin- 

 cipal section of the crystal cuts the ray : and as to the 

 extraordinary ray, it will not suffice to say that it never 

 resembles the ordinary ; w r e must say that it differs from 

 it as widely as possible ; that if the one, for example, is 

 coloured red, the other shows a bright green, and so on 

 for the rest of the prismatic tints. 



When this new kind of polarized rays are reflected 

 from a transparent mirror, we perceive other phenomena 

 not less curious. Let us conceive, in fact, to fix the ideas, 

 that one of these rays be vertical, and that it fall on a re- 

 flector of pure glass at an angle of about 35, this mirror 

 may be on the right side of the ray : and the inclination 

 remaining constant,- it may be turned to its left, before it, 

 or behind it, or in any intermediate position. We may 

 remember that the incident ray was white ; then, in any 

 of these positions of the glass reflector, the ray will not 

 have this colour : it will be now red, now orange, yellow, 

 green, blue, indigo, violet, according to the side on which 

 the glass presents itself to the incident ray ; it is, in fact, 

 precisely in this order that the tints succeed one another, 



both the images will be coloured, and their tints complementary. The 

 originally polarized light is divided again into two oppositely polarized 

 pencils in passing through the film, or as Professor J. Forbes has termed 

 it, ~Dipolariztd; others had termed it Depolarized. This is what Arago 

 here calls a new and entirely distinct kind of polarization; though the 

 term is, perhaps, not very happily applied. This is what was explained 

 at large in a previous note. 



