3$ OUTLINES OF NATURAL PHILOSOPHY. 



51. The terrestrial refraction found by means 

 of the preceding theorem, when the elevation is 



not very great, varies from - to ~- of the angle 



subtended by the horizontal distance of the ob- 

 jects ; and the radius of curvature of the ray, 

 therefore, varies from twice to twelve times the 

 radius othe earth. 



In the mean state of the atmosphere, the refraction is 

 about -7 of the horizontal angle, and the radius of 



purvature of the ray seven times the radius of the 

 earth. 



The terrestrial refraction must vary with the density 

 of the air, or with the barometer and thermometer. 

 The great differences, however, remarked above, 

 must be owing to some other cause. 



Jn the measurement of heights, the angle of elevation 

 should be diminished by one-fourteenth of the angle 

 corresponding to the horizontal distance, suppo- 

 sing the refraction to be of the mean quantity. 



52. The effect of refraction may also be allow- 

 ed for, by computing the correction of curva- 

 ture, as in 248, vol. i., and taking one-seventh 

 of it, for the number of feet, by which the ob- 

 ject is rendered by the refraction higher than it 

 ought to be. 



If 



