of REFRACTION. 247 



BxroE a ray coming from the bottom, and Tx^E a ray 

 coming from the top of the hill, failing upon the eye at 

 E, in a direction between thofe of the other two rays ; then 

 it is manifeft, that fuch a difpofition of the rays will produce 

 the obferved appearance. To erled; this, there muft have been 

 a very quick variation of the denfity of the air which lay be- 

 tween the two curves y v E, x w E, fo as to increafe the curva- 

 ture of the ray TxzE, after it cuts B w E in x, by which 

 means, the ray TxzE, might fall between the other two rays. 

 The phenomenon cannot be otherwife accounted for. As there 

 are not, that I know of, any records of a phenomenon of this 

 nature, the conflitution of the air muft have been fuch as but 

 very rarely happens, or fuch an appearance would before have 

 been taken notice of. 



The phenomena which I faw at the fame place, and which 

 I defcribed in the Philofophical Tranfactions of the Royal So- 

 ciety for the year 1798, I explained upon the fame principle, 

 that of a quick variation of denfity j and this was afterwards 

 confirmed by fome very ingenious experiments made by Dr 

 Wollaston. Perhaps this phenomenon may afterwards be 

 fubjeded to an experimental illuftration. 



VI. 



