34f OUTLINES OF NATURAL PHILOSOPHY. 



a considerable error in them will not produce 

 any sensible error in the refraction to be found. 



In conformity with this rule, the following method 

 of finding the horizontal refraction may be em- 

 ployed. 



46. A star which rises and sets due east and 

 west, would be as long under the horizon, as it 

 is above it, if there were no refraction. It will 

 be found, however, by observation, to be lon- 

 ger above than below the horizon : take half 

 the difference, and reduce it to degrees, or 

 parts of a degree ; then multiply this last by 

 the cosine of the latitude of the place ; the 

 product is the horizontal refraction. 



a. In this way, the mean horizontal refraction will 

 be found to be 33' 5 and here it would require a 

 great error in the latitude to produce an error at 

 all sensible in the refraction. 



b. This method requires a situation, where the stars 

 to be observed rise and set in the sea. It may, 

 however, be extended, though with less simplicity, 

 to all the stars, and to their descending, not be- 

 low the horizon, but below a given- altitude, or a. 

 given parallel to the horizon. 



47. Observe the altitude and azimuth of a 

 star of a known declination at the same instant j 



from 



