of the Atmosphere. 177 



with the angle of incidence, and becomes nothing at 180°, the 

 point opposite to the sun. 



At all these points the polarization is said to be vertical, being 

 in the vertical plane passing through the sun and the observer. 



In addition to the vertical polarization produced by the direct 

 illumination of the aerial particles, there must be an opposite 

 polarization by which the neutral points are produced. MM. 

 Arago and Babinet, and, we believe, every other writer on the 

 subject, have sought for this counterpolarization a in the secon- 

 dary illumination which the same aerial particles receive from 

 the reflexion of the rest of the atmosphere, which sends to them 

 light polarized horizontally"*, or oppositely to the light polarized 

 vertically. That is, all the phenomena of atmospherical polar- 

 ization are produced by the opposite action of two lights polar- 

 ized by reflexion, the one vertical, arising from the direct illumi- 

 nation of the aerial particles, and the other horizontal, produced 

 by a secondary illumination of the same particles by the rest of 

 the atmosphere. 



This theory of atmospherical polarization, omitting all con- 

 sideration of the light polarized by refraction, never appeared to 

 me satisfactory. There is no evidence whatever that such a 

 secondary reflexion exists, even in a perfectly cloudless sky, and 

 still less evidence that, if it did exist, it would be capable of 

 neutralizing the light polarized by reflexion at considerable dis- 

 tances from the antisolar point. It must be very feeble when 

 the neutral point is about to disappear at the close of twilight; 

 and as the polarization by direct reflexion must be visible when 

 the secondary reflexion ceases to be visible, this cessation ought 

 to be marked by a return of the neutral point to the antisolar 

 point, the place which it would occupy were there no secondary 

 reflexion. 



Were the neutral points produced by a secondary reflexion, 

 their distances from the antisolar point and from the sun ought 

 to be affected when the sky is more or less covered with clouds ; 

 but though I have observed the neutral point of Arago in a clear 

 part of the sky, I never observed that its distance from the 

 antisolar point was changed when the rest of the atmosphere was 

 obscured by clouds. 



On these grounds I was led to the opinion that the neutral 

 points must be produced by the opposite action of two polarized 

 lights which had nearly the same relative intensity ; and this 

 opinion was strengthened by observations which I had made on 

 the polarization of light by refraction and transmission through 

 piles of glass plates. In these experiments, published in the 

 Philosophical Transactions for 1814, I observed phenomena 

 * Babinet, Comptes Rendus, &c. 1846, vol. xxiii. p. 233. 



Phil. Mag. S. 4. Vol. 30. No. .202. Sept. 1865. N 



