gfei*' On the double Images cattfed by Atmofpkerical RefraBian. 



below this point I had no means of tneafuring with the acccuracy that would" be re- 

 quifite for determining the nature of the curves of dcnfity formed according to the 

 firll propofition. 



But the truth of the fecond propofition appeared capable of confirmation by experi- 

 nent; for the deviation of a ray is there faid to depend on the increment of denfity, and 

 lime of the ray's paflage, jointly ; accordingly, the deviation caufed by a given increment 

 ihould be in proportion to the extent of the medium. 



In order to try what efFe£l: a greater extent of medium would produce, 



Exper. 3. I made a redlangular glafs veflel, of which the fides were in the ratio of 10 

 to 30,6 ; and having put into it fome clear fyrup, with water on its furface, I meaAired 

 the greateft refra£tions through it in both diredions, and found them in the ratio of 10 

 to about 29. 



In another veflel, of which the fides were as 10 to 40,4, the refradions were, on 

 an average, as i to 4. 



Being now fully fatisfied of the effeft of different fluids, I made the following ex- 

 periment, whereby it appears, that the variations of denfity occafioned by difi^erence of 

 temperature between adjacent (Irata of the fame fluid, follow a fimilar law. 



• Exper. 4. Having put fome cold water into a fquare vefl^el, I covered its furface with 

 a piece of writing paper perforated with a few fmall holes, and then filled the veflel 

 cautioufly with boiling water. The paper nearly prevented any mixture of the hot and 

 cold water ; but, by floating gradually up, left them to communicate their heat by 

 contaft alone. 



"While they were in this ftate, I examined the appearance of remote objefts through 

 ihc varied medium, and found, that when my eye was removed four or five feet from 

 the vefl^el, the efl^efts were the fame as in the preceding experiments with difl^ercnt fluids ; 

 above any obje£t viewed through the cold water, I could difliinguifli two images of it, 

 the one inverted, the other ereft, as ufual ; but thefe appearances did not continue more 

 than five or fix minutes. 



Having thus eftabliflied, by experiments fufEciently varied, that the contiguity of two 

 fluids of unequal denfity is capable of occafioning all the appearances that have been ob- 

 ferved, I ftiall proceed to (hew by what means the air may be made to exhibit fimilar 

 phenomena. 



Exper. 5. I heated a common poker red hot, and held it fo as to look along the fide of 

 it at a paper 10 or 12 feet dift:ant. The rarefaftion occafioned by it caufed a perceptible 

 refradtion to the diftance of about ^ of an inch from the fide of the poker. A letter 

 feen more diftant from it appeared as ufual ; within that diftance there was a faint image 

 of it reverfed ; and ftill nearer to the poker was a fecond image dire<3:, and as diftindt 

 as the objeft itfelf, but fomewhat fmallcr, as in Fig 8, in which a fediion of the atmof-. 

 phere furrounding the poker is reprefented. At the bottom and fides it is nearly circular j 

 but upwards the circular form is loft in undulations, occafioned by the rapid afccnt of the 

 rarefied air. 



The 



