26 Prof. E. Edlund on the Path of Electrical Induction- and 



Two vessels, provided in the bottom with exit-tubes and stop- 

 cocks, were filled with mercury and placed so close together that, 

 when the cocks were opened, the mercury-streams were so near 

 each other that a spark could strike across. The mercury was 

 received in the two separate compartments of a glass vessel placed 

 below. When one vessel was connected with the point c by 

 means of a conducting-wire, and the other with the point e } de- 

 flections so distinct were obtained, as soon as the machine was set 

 to work, that there could be no doubt in reference to the elec- 

 tromotive property of the mercury in the above respect. 



The same experiments were repeated after the mercury had 

 been replaced by water containing sulphuric acid. Although a 

 slight spark, which was very distinct in the dark, passed between 

 the two jets, no evident deflections were observed which could 

 be ascribed to the disjunction-current. Notwithstanding this 

 negative result, I doubt not that water is an electromotor in this 

 respect, although the means at my disposal were not suitable to 

 show this, 



5. I have shown in a previous paper that, when induction-cur- 

 rents which result from voltaic induction have the opportunity 

 of traversing the spark of an electrical discharge, those currents 

 which tend to traverse the spark in the same direction as the 

 discharge can most easily penetrate it. The spark acts thus 

 like an electrical valve : of the two opposite induction-currents, 

 that passes in greatest proportion which is in the same direction 

 as the discharge. That the other also passes to some extent may 

 be seen from the appearance of the spark. When the induction- 

 coil is inserted in the circuit between e and g (see figure, p. 15) 

 the spark is duller, and cannot strike across such a distance 

 between the knobs as when the coil is removed — affording a proof 

 that the induction-current, which arises at the commencement 

 of the spark and goes in the opposite direction to the discharge, 

 does in fact partially traverse the spark. In this case, therefore, the 

 intensity of the spark is diminished by the induction ; its curve of 

 intensity is, as it were, flattened. When, on the contrary, the in- 

 duction-coil is placed between e and k, the intensity of the dis- 

 charge increases in consequence of the fact that the induction- 

 current formed at the commencement of the spark now traverses 

 it in the same direction as the discharge ; the spark has a greater 

 striking-distance than when the coil is removed ; its intensity is 

 increased. Now it might possibly be maintained that the reason 

 why the deflection of the disjunction-current is diminished by 

 the introduction of the induction-coil between e and g, or be- 

 tween e and k, is not that that induction -current which has the 

 same direction as the discharge traverses the spark in greatest 

 proportion, but that it is to be sought in the fact that the curve 



