172 Prof. E. Edlund on the Construction of 



Both disks were of the same size (that is, 50 millims. in diame- 

 ter) ; and the axis of rotation of the system, when suspended by 

 the cocoon-thread, went through the centre of each. The mag- 

 netic needle was let into a circular copper disk in such a manner 

 that the upper sides of the needle and of the disk lay in the same 

 plane, and their centres coincided. The disk and the needle 

 were soldered together so as to produce perfect conduction be- 

 tween them. The centre of this disk was 

 now made to coincide with the axis of ro- 

 tation, so that it became horizontal ; the 

 adjacent figure renders this arrangement 

 more intelligible, s is the silver mirror, 

 a b the round metal rod which constitutes 

 the axis of rotation, and c, d, and e are the 

 circular disks, in the latter of which the 

 magnetic needle is inserted. The coils 

 of the galvanometer surround the disk e 

 and pass between d and e, so that the 

 disks d and c and the mirror are at the top. 

 If, now, in the discharge electricity re- 

 mains upon any point, for instance at /, 



in the coils, it is clear that it can produce no rotation in con- 

 sequence of its influence on the disks c, d, or e. Of the electri- 

 city which is produced in the mirror s in consequence of induc- 

 tion, one part is repelled to the disk c, and the other attracted 

 to the disk d } and both thereby become innocuous. 



On testing, it was evident that this arrangement has a de- 

 cided advantage over that previously employed. In my former 

 experiments, a Leyden jar charged to saturation could not be dis- 

 charged through the galvanometer without producing a material 

 change in the position of equilibrium. When one of the galva- 

 nometer-wires was directly connected with one of the combs of 

 a Holtz's induction-machine, and the other ended with a knob in 

 the vicinity of the other comb, so that while the machine was at 

 work sparks sprang across, in my previous experiments a consi- 

 derable alteration was produced in the position of equilibrium 

 after the action of the machine had ceased. Hence to avoid this 

 a shunt was used between the conducting-wires, so that only a 

 small portion of the shock traversed the galvanometer. 



In the new arrangement of the moveable part of the galvano- 

 meter this bridge was quite superfluous, and the entire discharge 

 could pass through the galvanometer. There was indeed an 

 alteration in the position of equilibrium if the discharges were 

 particularly strong; but it was not so great as to act injuriously 

 on the accuracy of the measurements, and still less to render 

 them impossible. When one galvanometer-wire was connected 



