0n the double Images caufed hy Atmojphtrical RefraBion, 30 1 



Below the point </, where the curve terminates, vifion would be direft, fo than an 

 objed might be fituated fo a8 to be feen In all three ways at the fame time, dire£l at O, 

 inverted at I, and erc£t again at A. 



' I conCder the foregoing propofition as applicable to all cafes of varying denfity, whether 

 occafioned by mutual folution of different fluids, or partial rarefadion of the fame fluid j 

 and, by trial of various fluids, however different in denfity, or even in vifcidity, I find 

 that the refraSions obferve a law agreeable to the theory, as will appear by the following 

 experiments : 



Experiment I. Into a fquare phial containing a fmall quantity of clear fyrup, I put 

 about an equal quantity of water, in fuch a way that it floated on the furface of the 

 fyrup without mixing. For a fhort cimci the ftratum of union was fo thin that nothing 

 could be diftin£tiy feen through it. But foon, by mutual penetration of the water and 

 the fyrup, the effe£l:s reprefented at A, Fig. 6, were produced. 



Through the fyrup, a word written on a card, placed behind, was feen eredt, and 

 in its place ; through the adjacent variable medium, an inverted image was vifible 

 above the true, place} and alfo, above that, a fecond image of the fame objefi ap- 

 peared cre£t. 



When thefe appearances are firft difcernible, the variations of denfity are fo great, that 

 the objeft to be looked at may be in contafl with the phial ; but when the variations of 

 denfity become more gradual, and thereby the focus more diftant, any objeft fo near is 

 only elongated, and require to be removed an inch or two, to be feen inverted. 



Exper. a. Over the furface of the water, in the fame phial, I next put about the fame 

 meafure of re£lified fpirit of wine. 



At the ftratum where the water and fpirit united, the appearances were the fame ; but 

 fince the refra£live power of fpirit exceeds that of water, the true place of the obje£l 

 was feen uppermoftj the inverted and erett images are below^ Fig 6, B. 



When an oblique line d e r is viewed through any variable medium fo made, it appears 

 bent into different forms, according to Its iituation with refpeft to focal diftance. 



If it be at the diftance of the principal focus^ one point of It Is dilated Into a vertical 

 Jine, as / m. Fig. 7, A. 



If beyond that focus the portion I mis Inclined backwards, being an Inverted image of 

 d l; and m n Is another Image of the fame portion feen ere£t. Fig. 7, B. 



On this account, it becomes a convenient obje£t for afcertaining the ftate of any 

 medium under examination. 



In thefe experiments, the appearances continue many hours, even with fpirits of wine j 

 with fyrup, two or three days ; with acid of vitriol, four or five ; with a, folution of guni 

 arable, much longer \ but although their difpofition to unite is fo different, the ap- 

 pearances produced are neverthelefs the fame In all. 



The refracilon is greateft nearly in the plain of original contaft of the fluids, dimlnlfliing 

 from thence both upwards and downwards. The cxa£t rate of diminution above oi 



belovr 



