ICO 



It is worthy of notice that the second term is independent 

 of the thermometer, this circumstance enables us to put the 

 three last terms into a very convenient table, the a.rguments 

 of which are the zenith distance and height of the baro- 

 meter. 



12. The above expression for atmospheric refraction is en- 

 tirely independent on astronomical observations. 



The French tables are derived from observations of cir- 

 cumpolar stars. By these tables the refraction at 45° = 

 57",57 when the barometer shews 29j60 and Farenheit's ther- 

 mometer -50° Hence by equat. (1) art. 9« ; :,•!;; i^\;U; 

 57",57= ^^^^ (1-— ) = -^(0,99744). 



' ' • sill. i. ^ a ' sin.V \ '''•^ • / 



Therefore '4^ = 57^72. 



sin. \ 



Ey c25 observations of circumpolar stars made by myself 

 with the ei«ht feet astronomical circle (vid. art. 14.) I de- 



duce '^ = 57'',56. 



;j •;; sin. 1'/ ' ' 



• Thus the value of 'i=^ by the French tables is between the 

 values resulting from direct experiment and from my observa- 

 tions. 1 am inclined to give the preference to the result from 

 direct experiment for reasons afterwards mentioned. But the 

 difference between this result, and that from the French tables 

 is so small that no inconvenience can occur in adopting the 

 French tables. Thus, bar. 29,60 inches, and Farenheit's 

 therm. 60"- 



