INTRODUCTION. 63 



its lower side come to lie uppermost : the mean of 

 its vibrations in this situation gives the second 

 observation, which we will call a. 



2. After this, take out the needle, lay it on a 

 board, in which it is exactly fitted in a groove, 

 which shall protect the axis of the needle. Place 

 the north pole of a strong magnet on the middle 

 of the needle, and, with a pretty strong pressure, 

 pass it in a straight direction over the north end 

 of the needle : (not to get out of the straight di- 

 rection, it is advisable to fasten a ruler on the 

 board by the side of the needle :) this is repeated 

 several times, care being taken never to rub back- 

 wards from the ends of the needle to the centre, 

 or to slip beyond the axis to the other half of the 

 needle. In the same mannei', rub with the south 

 pole of the magnet, the south end of the needle ; 

 the poles will then be reversed. Whether it is 

 sufficiently magnetised appears, if, when it is laid 

 in the instrument, it assumes an inclination, which 

 cannot be increased by subsequent repeated rub- 

 bing. 



3. Repeat, with the needle magnetised to satu- 

 ration, and with the poles reversed, the experiment 

 described in N° 1, turning the graduated circle of 

 the instrument once to the east, and the other 

 time to the west. Let the mean of the one ob- 

 servation be called B, and of the other, b ; then, 

 if the four observations differ from each other 

 only a few degrees, the true dip is = I = A 4- a 



