WHICH REFLECT LtGHT. 325 



Tt is therefore proved to be impossible that the reflexions in 

 the atmosphere can be due to transparent masses on the surfaces 

 which are not parallel to each other, pushing even the indices of 

 refraction to an extreme limit. Nay, it might be proved that 

 indices which are so nearly = 1, are, for other reasons, not at all 

 admissible, as they would have peculiar phaenomena as conse- 

 quences which would be quite at variance with the reality. 



Summing up the results of the foregoing investigation, we 

 find that if we do not assume that the reflexion takes place on 

 opake bodies floating in the atmosphere, then along with the 

 reflexion refraction must also take place ; and that this refrac- 

 tion under all circumstances would produce phaenomena com- 

 pletely contradictory to our commonest observations, with the 

 exception of the single case, that the reflecting bodies are thin 

 plates w\i\i parallel surfaces. We, hence, almost of necessity, are 

 driven to the assumption of fine vesicles of water, which even in 

 serene weather float in the air and produce the reflexion. 



[J. T.] 



