228 On Illumination and Polarization in Transparent Substances. 



VIII. I have hitherto always supposed that the corpuscles 

 were spherical. Is this condition necessary for the explanation 

 which I have given of the phenomena of illumination ? I think 

 not. For, on the one hand, whatever be the shape of the parti- 

 cles, the surface is always tarnished, and the phenomena of in- 

 terference which I have pointed out will always be partially pro- 

 duced. On the other hand, we have not to consider a single par- 

 ticle. In a layer perpendicular to the luminous pencil there is 

 a large number of particles ; there must, then, in general be 

 in this layer as many surfaces in a certain position as in the 

 position symmetrical with it. These symmetrical surfaces will 

 produce the interferences of certain components of the reflected 

 vibrations, as if they belonged to the same particle. 



This is the summary of my theory, or rather of my attempt at 

 a theory. 



Although this letter is already long, will you permit me to 

 add a few words? You will have seen that M. Lallemand 

 has answered {Comptes Rendus, Dec. 20) the note I communi- 

 cated to the Academy (Comptes Rendus, Dec. 6), and that he 

 still thinks the illumination of liquids and of solids should be 

 attributed to a lateral propagation of the motion of the aether, 

 and not to particles diffused in the transparent medium. The 

 reasons which he adduces in support of his opinion do not seem 

 to me conclusive; but I will not dwell upon this subject. I 

 will merely mention that I have recently made a series of ex- 

 periments by taking water and producing light precipitates in it 

 by means of certain reagents. Whenever the precipitate was 

 formed of particles so fine that they were not visible to the 

 naked eye, and in such small quantity that the transparency of 

 the water was not sensibly diminished in diffused light, I ob- 

 served a great increase in the illuminating power of the line of 

 light, with the same phenomena of polarization. 



Thus I take water from the Lake of Geneva which has quite 

 settled, and which has very little illuminating power; it con- 

 tains some traces of chlorides in solution; I add to it, by means 

 of a glass rod, a drop of an extremely dilute solution of nitrate 

 of silver. The line of light becomes far more visible, and the 

 polarization is complete. I add a larger quantity of nitrate of 

 silver, and the water becomes a little dull ; the line is still more 

 distinctly visible, of a bluish tint ; but if we examine with a NicoFs 

 prism, in a suitable position, the extinction is no longer complete ; 

 the line is still seen, and is of a beautiful deep blue (almost the 

 tint of the indigo of the spectrum). 



It appears to me probable that this incomplete polarization is 

 due, at least partially, to multiple reflections when the particles 

 in suspension become very numerous. If water containing 



