I 



through Rarefied Air. 325 



particles in the space between the electrodes. This space will 

 consequently conduct electricity better^ and the contact current 

 can pass as a glimmering discharge. That for the transmission 

 of this current a perfect conductor is not necessary, is proved by 

 the fact that the same can be transmitted through a long strip 

 of tolerably moistened paper. These premises being granted, 

 the observed actions of the induction current can be deduced 

 without difficulty. When the whole current of the induction 

 apparatus is passed through well-rai-efied air, between a small 

 and a large surface, and the direction of the interruption current 

 is from the small to the large surface, this current alone is trans- 

 mitted, and that as a glimmering discharge : in the discharge 

 circuit, therefore, only a small amount of heat will be generated, 

 a galvanometer will be strongly and constantly deflected in one 

 direction, and an electrolyzable substance will be decomposed in 

 the ordinary manner, so that u certain constituent of the same will 

 be separated at a certain electrode. If the direction of the current 

 be reversed, so that the interruption current pass through the 

 rarefied air from the great to the small surface, this current will 

 suffer partial discontinuous discharge ; hence a greater heat will 

 be produced in the discharge circuit, the deflection in the galva- 

 nometer will be less, and a smaller quantity of the electrolyte 

 will be decomposed. These actions will be modified by the 

 glimmering discharge of the contact current, which can now 

 pass in consequence of the greater conductibility of the space, 

 its heating action will be weak, the deflection and decomposition 

 which it causes will both be strong, and in direction opposed to 

 the corresponding actions of the interruption current. The de- 

 flection of the needle, therefore, can be still further diminished, 

 or destroyed, or even reversed; the separation of a particular 

 constituent at the former electrode may be prevented, or it may 

 even be disengaged at the opposite electrode ; or, lastly, the 

 effects of both currents may remain perceptible : the constituent 

 in question may be separated at both electrodes, and the needle 

 may be deflected sometimes on one side, sometimes on the other. 

 In fact, tliis changeability in the result takes place, not only in 

 different experiments, but frequently in the same experiment ; 

 nor can it surj)rise one who reflects that it is produced by the 

 changeable successive action of two o])posite currents. On the 

 contrary, with the former direction of the induction current, the 

 result, being influenced by one current only, is always essentially 

 the same. 



Though not immediately connected with the subject in hand, 

 I may here remind the reader of a paradoxical fact upon which 

 some light may, perhaps, be thrown by the above-proved influence 

 of the interruption current upon the passage of the contact 



