222 M. L. Soret on Illumination and Polarization 



be the same as that given by M. Lallemand, except in what 

 concerns the origin of these waves. 



I confine myself here to mentioning this first hypothesis : as it 

 has not hitherto been supported by any experimental proof, nor 

 even by any analogy with other phenomena of light, it would be 

 useless to discuss it at present. 



II. A second hypothesis would assume that a phenomenon of 

 diffraction is produced analogous to that of gratings. Let us 

 imagine the transparent medium divided into layers perpendi- 

 cular to the pencil of light ; in each of these layers there are 

 particles which should act like a very fine powder deposited 

 on a transparent plate ; that is to say, part of the light is no 

 longer propagated in a right line, but diffuses itself in various 

 directions. The vibrations producing this scattered light re- 

 tain the form and the direction of the vibrations of the illu- 

 minating pencil ; hence, if this is polarized, the scattered light 

 will consist of rectilinear vibrations parallel to those of the pencil. 



I am inclined to think that this diffraction really takes place, 

 and contributes to the production of the phenomena observed, 

 especially in directions which differ little from the direction of 

 the illuminating pencil ; but in any case the particles should 

 also reflect light, and this effect ought to predominate over that 

 which has been mentioned. 



III. Let us, then, consider the reflected light. Let abed (fig. 1) 

 be a body illuminated by a pencil of parallel rays, s' b, s a, s f ' c. 



Fig. 1. 



Suppose the observer's eye placed in any position o, looking at 

 the body abed. For shortness' sake let us call the angle of 

 vision the angle o a s formed by the visual ray and the inci- 



