104 



(1) To shew the effects of the abovementioned irregularity 

 ot refraction, by which it appears that at zenith distances not 

 greater even than 80°, no use can be made of observations for 

 the nicer purposes of astronomy. 



(2) As it is reasonable to suppose this unexplained irregu- 

 lariTiy* will disappear from a mean of a great number of ob- 

 servations, this star, which is just at the limit where the 

 quantity of refraction ceases to be independent of the vari- 

 ation of density, may also serve as a criterion of the exactness 

 of the value of "'"~;. or of the quantity of mean refraction. 



The refraction observed and the refraction computed by the 

 formula in Art. 11. are placed by the side of each other, and 

 also the correction of the computed refraction to give the ob- 

 served refraction. This correction is often far beyond the limit 

 of the error of observation, and is to be attributed to the above- 

 mentioned irregularity of refraction. 



* The hypothesis upon which refractions are computed is that the different strata of air 

 / are concentrical with the earth's surface, circumEtances may he easily imagined to affect 

 this hypothesis, with respect to low stars. 



