— 8 — 



longitudinal sections of the bar. The cross section was detormined by means 

 of a screw micrometor, ton readings at difforent portions of the bar giving a 

 fiiir average. 



For determining the time of oscillation, the magnet was sot in vibration 

 in a wooclen caso with ftmr glass sides, one of which could bo opened at plea- 

 sure. Tiie magnet was suspended_by two loops of silk fibre, whidi were united 

 into a single fibro of the same material at a distance of about 5 cm. from the 

 magnet, the length of the single fibro boijig about 20 cm. To bring the 

 magnet Luto tho horizontiil position, tho floor of the caso was first luveiled by 

 means of three levelling screws belonging to tho caso : the magnet was then 

 lowered close to the floor and was made parallel thereto by sliding atljusfmeut 

 of one of tho silk loops. The bar was now raised by winding the suspending 

 fibro at the fop of tho caso. 



When the bar was settled in its adjusted position, it was sot in vibration 

 by bringing a piece of iron outside the case at nearly the same level as tho 

 bar. Tho oscillation was observed by sighting a reflecte<l image of a »calo 

 placed at a distance of 50 cm. from the magnet, whoso polished end was 

 used as tho reflector. One lino in tho scale was marked, and when tho 

 image of that lino passed the wire of tho toloscope tho observer presse«! tho 

 break-circuit-key, which was in connection with tho chronograph. From ten 

 to twenty successive signals were thus given and the magnet was left to vibrato 

 for about five minutes, when another series of ten to twenty signals was 

 made. From the ten successive marks in the chronograph sheet the time of a 

 single oscillation was roughly determined, und the number of oscillations in five 

 minutes was inferred as in the pendulum experiments. 



The determination of tt was by tho tangent method. Tho magnetometer, 

 of tho ordinary small mirror reflecting form, was set upon a tripod stand. Tho 

 deflecting magnet slid along a groove cut in tho upper surface of a brass ro<l, 

 which was specially constructed to suit tho apparatus (see Fig. 1). This brass 

 rod rested on the tclescopo supports of a theodolite stand, which was truly 

 centred with the magnetometer tripod, but ha<l no contact therewith. Tho 

 lino of supports was adjusted to the direction of magnetic east and west by 

 an electro-magnetic method, which will bo describi-d below in the account of 

 the Declination Experiments. When this adjustment was effected, tho brass 

 rod was laid in position. Through a circular hole cut out from the centre of 

 the rod, the magnetometer passed; and tho mirror with its attached magnets 

 was carefully adjusted to the proper level. Tho \'-groore in the l)rass r<^l 

 was graduated from the centre in Ijoth directions. The bar maguet was 

 mountcil on this V at two distances r, and r.^, whose ratio was approxi- 

 mately 1 : 1.32, this _being according to Maxwell the best ratio to take. 



To measure tho angle of deflection a wooden arc of radius 85 cm. was 

 graduated to minutes and was placed to one side of tho theodolite on 

 a wooden tripofi support. Tho reflected image of tho scalo was obsorvud 



