ON THE POLARISATION OF THE ATMOSPHERE. 239 



sidering the congeries of separate molecules which constitutes the atmosphere as 

 a rough or dispersing surface, I made a series of experiments on such surfaces 

 as were likely to act upon light in the same manner as the aerial molecules. In 

 these experiments, which have been already communicated to the Society,* I 

 found that such surfaces not only polarised partially the incident light, but pro- 

 duced neutral points, like those of the atmosphere, by the opposite action of 

 the rays which they polarised by reflexion and refraction. 



* See page 205. 



VOL. XXIII. PART If. 3 U 



