494 WE15ER OV THE MEASUREMENT 



cylinder itself, B, with the prong and the screw e e, can also be 

 screwed up or down by means of the nni,ff; and thus the 

 vibrating reel may be arranged in the proper position as regards 

 the multiplier, in the centre of which it should oscillate. At 

 the same time the roller B, which is moveable in the prong e e 

 around the peg m, assumes a state of equilibrium as soon as 

 the vibrating reel is suspended freely from the bifilar metaUic 

 wires, since these wires act at b and b' as it were at the ends of a 

 lever, the centre of motion of which is at m. Thus the load of the 

 vibrating cylinder is equally divided between the two threads. 



To allow of the approximation or separation of the two bifilar 

 wires, the rollers a and a' are set in broad prongs, which, as seen 

 in the figure, terminate in screws, by means of which they can 

 be approximated or separated between two metallic plates (indi- 

 cated by the lines engraved perpendicularly) with the nuts c c 

 and c' c'. The latter are fitted into a kind of case, indicated in 

 the figure by lines drawn obliquely, in which they are fixed by a 

 peg, but are not impeded as regards their rotation. The roller a, 

 with its prong and screw, plate and nut c c, is isolated from the 

 roller a', with its prong and screw, plate and nut c' c', because the 

 circular discs d d and d' d', which are perforated in the centre, and 

 which connect them above and below, are made of ivory. To 

 allow of the bifilar metallic wires being brought out conveniently, 

 the nuts c c and c' c' terminate in trumpet-shaped projections, as 

 shown in the figure, from which hangs a wire ff g and^y thrice 

 wound round. Hence a galvanic current takes the following 

 course : — If it enters atff, it ascends to ff, is communicated to the 

 nut c c, and the roller a (it also ascends to b, but as b is isolated 

 it returns), and runs down the threads to o; from o it proceeds 

 (fig. 2) further down through the centre J!J of the transverse beam, 

 then to its extremity r, where by the metallic contact with the 

 support it runs down it, and at m enters the extremity of the reel 

 itself, through the coils of which it continues, again making its 

 exit at d, but again passing to the other support at m' through 

 k, from r' along the transverse beam to its centre, and from this 

 up to o' ; from o' the ciu-rent (fig. 5) again runs over the other 

 roller a' into thu nut c' c', and finally arrives at the other con- 

 ducting wire, g' (/'. Thus the current, to arrive at one conduct- 

 ing wire g'ff' from the other gg, must necessarily run through 

 the vibrating reel, inasmuch- as the wire from,^' to g is perfectly 

 isolated. To do away with the torsion of the bifilar metallic 

 wires, the whole of the upper portion of the instrument as far as 



