MAGKETIC SURVEY OF JAPAN. 185 



one-fifth its breadth. The.se dimensionsi were carefully cah.-ulatcd * 

 so as to make the error due to an excentrie j)0>>ition of the mag- 

 netometer a minimum. 



To bring the coil into adju.stment, it i.s neces.sary to operate a.s 

 follows. Place the stage and magnetometer on the theodolite, and 

 mount the coil witli it.s pivots resting on the Y's (PInte VI, fig. 1). 

 Adjust the Y's into an approximate ea.st and west direction by sight- 

 ing the freely hanging mirror edge-on through the jâvot cores. Lay 

 the coil horizontal, so that the ends of the coil now face north and 

 south. On one of these ends, which is partly closed, a small pin-hole 

 is made accurately at the centre. On the other end, exactly opposite 

 it, a wire (Ji in the cut) is stretched. When this wire is sighted 

 through the \^\u-\h)]e -.md throuyJi tJic small hollow tiiagnef, then we know 

 that the inao-net is truly centred with reu'ard to the c<jil. If the wire 

 cannot so be .seen, the ma^'iietometer must be adjusted in no.^ition bv 

 means of the screws pro\ ided in the ring .socket in which the base of 

 the magnetometer rests. If every thing is in accurate adjustment, ;i 

 reversal of the coil, so that what was at first the east pivot becomes 

 west and vice cersa, will still leave the pin-hole, magnet, and wire in 

 one and the same line. If it is not so, then the j)in-hole and wire are 

 at faidt. It is evident, however, that once the adjustment i.s made, 

 it is made once for all as fir as the coil is concerned — that is until the 

 wire should oive wav. The maii'netometer bein«»' thus ceiiired east 

 and west, it must next be centred north and south. Tiiis is done 

 with sufiicient accuracy bv sitrhtin'j" the mao-net throuuh the hollow 

 pivot of the coil, and working the north-and-south adjusting screws 

 until the magnet hangs central. 



If the small telesco])e with attached scale and lamp-counterpoise 



* The details of the calcnlation are given in the Fa^ev in the Proceediiig.-i of the liotftil 

 S'jcicty of Edhibiinjii (1884-6) already referred to. 



