Disjunction-currents through Gases of various Densities, 



15 



copper wire de terminates in a 

 brass knob in the vicinity of b. 

 From c and e insulated conduct- 

 ing wires pass to the knobs / 

 and#. Two other conducting 

 wires pass from the joints c and 

 e to the galvanometer #, of which 

 instrument I have already given 

 a description. The bridge h, 

 which consists of a German- 

 silver wire, connects the points 

 % and k with one another ; / is a 

 conducting wire leading to the 

 earth, and m a rheostat consist- 

 ing of fine German-silver wire. 

 The deflections indicated by the 

 needle of the galvanometer when 

 the machine is in action are pro- 

 duced by three different cur- 

 rents — namely, (1) that part of 

 the discharge of the machine 

 which passes through the coil of the galvanometer, (2) the dis- 

 junction-current formed in the spark between / and g, and (3) 

 the induction-currents which are produced by the discharge 

 in the coil of the galvanometer. As regards the first current, it 

 is under ordinary circumstances so inconsiderable as compared 

 with the others that it need not be taken into account. The 

 disjunction-current, on the contrary, produces a considerable de- 

 flection ; but this is very materially diminished when the induc- 

 tion-currents from the galvanometer traverse the spark between 

 f&n&g. As I have already shown in a previous paper*, the 

 spark acts like an electrical valve ; that is, it allows one of the 

 two equal but oppositely directed induction-currents to pass in 

 larger quantity than the other. The difference in the two in- 

 duction-currents thus occasioned acts therefore on the magnetic 

 needle, and always in such a manner that the deflection produced 

 by the disjunction-current is diminished. When, therefore, the 

 magnitude of the disjunction-currents under various circumstances 

 is to be investigated, the experiments must be so arranged that 

 the action of the induction-currents on the magnetic needle shall 

 be nullified as much as possible. This may be most easily effected 

 by inserting a bridge between i and k. In this case only part of the 

 discharge goes through the coil, and the induction is therefore 



* Oefo. af Vet. Ac. Forh. 1868, p. 457 ; Poggendorff's Annalen, vol. 

 cxxxvi. p. 337 ; Phil. Mag. S. 4. vol. xxxvii. p. 41. 



