10 HISTORT of the SOCIETK 



The only circumflance in which the appearances did not per- 

 fecflly correfpond with this hypothefis, was, that the two rain- 

 bows did not interfe(5t one another in the horizon, but rather a 

 little above it. This, however, ought to have no great weight, 

 as the reflc<5led image of the fun cannot have prefented to the 

 cloud a difk fo regular and well defined as the fun itfelf and 

 the accidental rainbow muft have fomewhat participated of this 

 indiftindlnefs. 



When phenomena of this kind occur, it would afford a fure 

 means of trying the juflnefs of the explanation if the in- 

 clination of the two bows were obferved, and alfo the fun's 

 altitude at the fame time. Thefe two things are neceffarily 

 connedled ; for if we call I the angle of their interfedlion, E 

 the elevation of the fun, and S the angle fubtended at the eye 

 by the femidiameter of the rainbow, if complete, an angle which 

 is conftantly the fame, and nearly equal to 42", it is eafy to 



infer from fpherical trigonometry, that fin ^ I =: SiTs" 



Computing from this formula, the inclination of the two 

 bows in the prefent inftance comes out nearly 5 ° ; fomewhat 

 greater than I was inclined to eftimate it by the eye. 



Phenomena of this kind can but rarely occur, as the necef- 

 fary conditions will not often come together. The principal 

 rainbow mull be over the fea ; the fea itfelf muft extend fome- 

 what on the fide toward the fun ; it muft be fmooth and 

 tranquil, and the fun fo low that the light refledled from the 

 water may be confiderable. Were it ever to happen that the 

 accidental bow was completely formed, the effed could not fail- 

 to be very ftriking. 



CHEMISTRY. 



