On the double Images caufed by Almofpheticd RefraS'ion. 307 



. The following experiment fliews that it is not to be expected but by very great extent 

 of furface. 



Exper. 1 2. I placed a dark coloured board in the fun fliine, and having examined the 

 refradion along its furface, I made a wet line along it, with a fpongc dipped in boiling 

 water. Notwithftanding this additional heat, the refradion, in the diredion of the wet 

 line, was far lefs than over the reft of the board, although I took care to obfervc the effed 

 before the furface could be cooled again by evaporation. 



I (hould therefore exped the depreflion of the horizon at fea where the refradion 

 occafioned by heat muft always be counteraded by evaporation, never to exceed a few 

 minutes ; and that any one in a fituation commanding a view of the fea, by attention to 

 the various degrees of the dip of the horizon under different circumftances, might foon 

 form fome eftimate of the proper allowance to be made for brightnefs of the fun at the 

 time of an aftronomical obfervation, or for difference of temperature between the fea 

 and air. 



Having now examined the feveral peculiarities of refradion which I propofed for con- 

 (ideration, I fliall, in few words, recapitulate the purport of the foregoing pages. 



According to the theory here given, there appear to be two oppofue ftates of the 

 atmofphere, either of which may occafion objeds to be feen doubled or tripled, fince both 

 increafe and decreafe of its denfity, when partial, produce the fame efFeds. 



It has been explained, 



I ft. Why air heated by the moderate warmth of the fun's rays occafions objedis ta 

 appear doubled and inverted. 



adiy. Why rarefadion, by a higher degree of heat, gives an additional image which 

 is not inverted. 



3dly. In what ftate of evaporation the increafe of the air's denfity brings diftant objeds 

 into view by unufual elevation. 



4thly. Under what circumftances evaporation may alfo produce an inverted image 

 lefs elevated. 



And it is probable that the fame reafoning will afford a ready explanation to 

 other varieties of terreftrial refradion, that may have been, or may hereafter, be 

 obferved. 



R r 2 J Memoir 



