482 Intelligence and Miscellaneous Articles. 



cuit that gives rise to this sound ; the closing produces no effect, at 

 least in my apparatus. 



M. de la Rive, in 1843, discovered that a current interrupted by 

 means of a rheotome produces a sound in the iron of an electromag- 

 net ; but I think that the phenomenon I have just described has 

 not been previously signalized. 



Listening for a sound in the nucleus of an electromagnet may be 

 regarded as a new process of investigation ; we have seen that it 

 reveals the intermission of the current in circumstances in which 

 the known methods are insufficient. I will mention an instance 

 which shows that this method accords with the others when they 

 can be employed simultaneously. 



Experiment 2. — "When the spark from the rupture of the prece- 

 ding circuit is observed with the aid of the revolving disk, after the 

 manner made known by me to the Academy on the 7th of April 

 last, the spark appears compound. When it bursts in alcohol, and 

 the platinum point and the mercury are connected with the arma- 

 tures of the condenser, the revolving disk shows that the spark is 

 composed of four or five successive bright strokes ; the intervals 

 between the strokes diminish from the first onwards. (I have 

 already described the division of this spark by another process, 

 Bulletin de la Societe Philomathique, 13th May, 1865 ; and the 

 journal Vlnstitut, 31st May, 1865.) 



The sound produced by the rupture-spark presents a similar 

 mode of division ; it is the same with that which is heard in the 

 condenser, and with that which takes place in the nucleus : these 

 three sounds are composed exactly in the same manner. 



The production of a sound in the condenser proves that there is 

 a partial discharge through the insulating substance, although at no 

 part does this appear to be pierced. 



The sound of the condenser is augmented by increasing its sur- 

 face, up to a certain limit which cannot be exceeded. At the same 

 time the spark between the mercury and the platinum point is seen 

 to diminish. These modifications indicate a change in the distri- 

 bution of the electricity — which could be mathematically analyzed 

 by regarding the alcohol of the interruptor as the insulating plate 

 of a second condenser united to the first by the armatures of the 

 same sign. 



I think there is no essential difference between the intermission 

 of the current which accompanies the breaking of the circuit in the 

 circumstances I have just described, and that presented by my first 

 experiment. 



Experiment 3. — The platinum point of the mercury interruptor 

 is screwed into a fixed nut, so that it can be raised or lowered. The 

 mercury and the point communicate respectively with the armatures 

 of a condenser of 1 square metre surface. The rest of the circuit 

 is arranged as before. 



The point being dipped into the mercury, is gradually raised until 



