direct induced Current between Electrodes of the same kind. \ 09 



inverts or reestablishes the connexion of each end of the covered 

 wire with a given electrode. The similarity of these electrodes 

 is such that no effect is produced. A more simple method is to 

 change directly the terminal electrodes without varying their 

 distance. The success of this experiment requires the employ- 

 ment of similar and well-isolated conductors for connecting the 

 ends of the covered wire with the arms of the discharger. The 

 outer circuit should be free from any roughness by which the 

 charge might be dissipated in the air before traversing the in- 

 terval between the electrodes. 



Thus, in air, between exactly similar electrodes kept at a con- 

 stant distance, the spark passes abundantly or ceases to pass, 

 according as one of them is or is not positive. The spark 

 passes when this electrode is connected with the outer positive 

 end of the induced circuit*. I have verified the fact with a bob- 

 bin of larger size, furnished with an independent Foucault's 

 break. If the direction of the voltaic current is interrupted very 

 suddenly, a small spark accompanied by a peculiar sound still 

 passes, but is not usually followed by another. At the same 

 time the break changes the course and gives rise to a sound cha- 

 racteristic of the rupture of the induced current. 



This interruption is not destroyed when hot and moist air is 

 blown into the interval to be passed. If a moderate conductor, 

 such as a wooden rod, be brought near the negative electrode, 

 the electric distribution of the flashing spark is modified f . But 

 this modification only lasts as long as the conductor is present. 



It has been long since noticed that in non-partitioned ma- 

 chines the electricity does not accumulate in the same manner 

 at the two ends of the induced wire. The induced wire and 

 the central iron core condense the fluid upon the inner end, 

 which does not give off a spark upon a neigbouring non-isolated 

 conductor, as is the case with the outer end. But this pecu- 

 liarity is independent of the fact that this end is positive or 

 negative; and it is considerably weakened in the partitioned 

 machines, simply by their construction ; it appears nevertheless 

 that the tension remains slightly preponderating in the outer 

 knob. When it is positive, as the negative knob becomes the 

 seat of an elevation of temperature, this twofold circumstance 

 determines the discharge, according to a mechanism analogous 

 to that of Sullin's card-perforator, of which M. Riess has given 

 the explanationj. A first partial discharge sets in motion the 



* By outer end in a partitioned machine, I mean that which corresponds 

 to the outer termination of the induced circuit of a non-partitioned bobbin 

 when the spark is exchanged between a point and a plate. 



t This property has already been studied by M. Riess. See Poggendorff's 

 Annalen, vol. xcix. p. 637 (1856). 



X Op. cit. vol ii. p. 213. 



