MAGNETISM 



Considering this error to exist alone, if Z, Fig. 228, i> t lu / 

 the first position, we read NX instead of MI. In the second 

 ZZ' is brought to the position Z'"Z" and we read NZ". Hut X is as 



Fio. 227. 



much below H as Z is above it, so that NZ + NZ" = 2NH, and 

 the mean of the readings eliminates the zero error. 



The a\i- of figure of the needle does not in practice coincide 

 exactly with the magnetic axis. To eliminate the error 



Fio. 228. 



produced the needle is taken up and its face reversed. The 

 positions of the ends are again read. Then, as with the declination 

 needle, it is easy to see that the mean of the readings, with tin 

 face of the needle West and with its face East, will give the 

 position of the magnetic axis. At the same time part of tin 

 error due to the centre of gravity lying out of the axis of rotation 

 will be very nearly corrected. For if G, Fig. 229, is the cc i 

 of gravity and O the axis of rotation, drawing the rectangle ( )AGB, 

 the moment of the weight of the needle acting through G will be equal 

 to the moments of two equal weights at A and B. By turning the 

 needle over, the moment of the weight at A is reversed, and tin 



