25 S FEAGMENTS OF SCIENCE. 



phere and a hot summer day, of unevaporated particles of 

 water." But a polarizing angle of 45° corresponds to a 

 refractive index of 1 ; this means that there is no refraction 

 at all, in which case we ought to have no reflection. Brew- 

 ster, therefore, and others came to the 'conclusion that the 

 reflection was from the particles of air themselves. Dr. 

 Rubenson, of Upsala, made the angle enclosed between the 

 direct and reflected beams 90° 2' ; " the half of which," says 

 Mr. Buchan, in his excellent little " Handy Book of Me- 

 teorology," " is so near the polarizing angle of air as to leave 

 no doubt that the light of the sky, as first stated by Brew- 

 ster, is polarized by reflection from the particles of air." 



If you doubt the wisdom, acknowledge, at all events, 

 the faith in your capacity which has caused me to bring 

 so entangled a subject before you. I would fain believe, 

 however, that even the intellect which draws its culture 

 from a totally different source, may have its interest excited 

 in subjects like the present, dark and difficult though they 

 seem. I do not expect that you will grasp all the details 

 of this discussion ; but everybody present will, I think, see 

 the extremely important part hitherto played by the law 

 of Brewster in speculations as to the color and polarization 

 of the sky. Let me now endeavor to demonstrate in your 

 presence, firstly, and in confirmation of our former experi- 

 ments, that sky-blue may be produced by exceedingly mi- 

 nute particles of any kind of matter ; secondly, that polari- 

 zation identical with that of the sky is produced by such 

 particles ; and thirdly, that matter in this fine state of di- 

 vision, where its particles are small in comparison with the 

 height and span of a wave of light, releases itself completely 

 from the law of Brewster ; the direction of maximum polari- 

 zation being absolutely independent of the polarizing angle 

 as hitherto defined. 



Into this experimental tube, in the manner already de- 

 scribed, I introduce a vapor which is decomposable by the 



