80<S On the double Images caufed by Almofpkerical Rejra^ion. 



at e, as a point, through b and d, would be made to diverge to a and c ; confequently, vifual 

 rays could not, under thefe circumftances, interfe£t each other, and no obje£ts could 

 appear inverted.- 



But whenever the loweft ftrata of the air become faturated with moifture, the variations 

 between the faturated ftratum, and the incumbent atmofphere of the common denfity, 

 will follow a law fimilar to what is found at the confines of other fluids of unequal 

 denfity ; hence, iiiverfion will become vifible, as there will be a point below which the 

 increment of denfity will decreafe, and where the refraftions will confequently be lefs, 

 although through a denfer medium. 



Exper. to. To produce thefe appearances, I procured a trough of thin deal, five feet 

 long, one inch wide, with fides 2i inches high, and clofed the extremities of it with 

 glafs. A fedlion of it is given in Fig. lo. 



When the bottom was wetted with ether, the greateft refra£lion was, at intervals, 

 more than -^ of an inch from the bottom of the trough ; and, beneath this height, I 

 faw a fecond image inverted, when my eye was removed to 14 or 15 feet diflance, and 

 the objeft was about 70 feet. The focus feemed at the fame time to be about nine feet 

 diftant. 



There was not depth enough of uniformly faturated atmofphere for the objedt itfelf 

 to be feen at through it, but its true place, compared with that of the images, is re- 

 prefented at a. 



Exper. II. When I made ufe of rectified fpirit in the fame apparatus, I had alfo 

 fuIHcient proof that the laws of evaporation would admit of fuch appearances being pro- 

 duced ; for the fame obje£l now appeared curved downwards, as in Fig. 1 1, fo that rays 

 nearer to the bottom were manlfeftly lefs refracted than fuch as paffed at fome difliance 

 above. A degree of convergency muft therefore have been produced, although the 

 diftance at which the rays would meet was beyond tliat of my eye, and circumftances 

 would not admit of my removing beyond 35 feet. 



The evaporation of water could not be expe£led to produce any fenfible efFetl: of this 

 kind in fo fliort a fpace ; but in a view of fome miles extent, there can be no doubt, from 

 the foregoing experiments, that evaporation from the furface of the fea, in fuch a ftate of 

 the atmofphere as would allow the lower ftrata to be faturated, is capable of occafioning 

 all the phenomena which have been defcribed, and probably was the caufe of thofe which 

 Mr. Vince obferved. 



Since heat alone tends to deprefs objects, and evaporation produces apparent elevation, 

 it is probable, that in the inftance of refraQion related by Mr. Dalby (Phil. Tranf; for 

 i795» P- 587) '^^ ^^^' °^ '^^ ^"" ^^^ '^^ principal agent, and that the moifture rather 

 tended to countera£l than aflift its adtion. 



Simple inverfion may generally b; feen when the fun (hines upon a dry even road of 

 ^ or 4- niile extent ; but when the ground has bcca wet, I have very rarely feen it, and. 

 have even failed of difcerning it, when the heat has been fufficicnl to raife a ftream from 

 the ground. 



4 'The 



