222 



PROCEEDINGS OF THE AMERICAN ACADEMY 



All these results were taken, interrupting the primary cirQuit by 

 separating two copper wires held one in either hand. Practice made 

 the results then obtained quite accurate. I at first tried breaking from 

 the surface of mercury ; but, beside the irregularity in the amount of the 

 deflection, there was a most unaccountable change of polarity every few 

 moments. I distrusted the evidence of my own senses so much in this 

 case that I asked several other peojile to observe for me, without pre- 

 viously telling them any thing of this change of polarity. In each 

 case, the observer noticed the reversal for himself. The deflection 

 was almost always in one direction when the circuit was broken 

 between the poles, and in the other direction when the circuits were 

 broken outside of the poles. This rule is not, however, absolute. 

 This reversal of polarity only occurred when copper was used to break 

 from the surface of the mercury. The separation of iron from iron, 

 or copper from copper, or iron from mercury, never gave any reversal. 

 This reversal was best seen with the electrometer. 



At several different times, I took a series of observations upon the 

 deflection, by breaking inside the poles and outside of them when the 

 distance over which the spark had to leap was varied. The curves 

 obtained by laying down these deflections were not, as one might sup- 

 pose, hyperbolas, but were apparently exponential curves, having the 

 axis of X as an asymptote, but not the axis of y. A series of very 

 careful observations was taken by observing the deflections when the 

 distances over which the spark passed were small. The curves ob- 

 tained by breaking the primary between the poles were similar to those 

 obtained when the primary was broken outside. Tlie observations 

 were taken by breaking the primary with the interrupter of one of 

 Ritchie's induction coils. The sparks passed between two circular 

 discs of copper 10 cm. in diameter. In the centre of one of the discs 

 was an almost imperceptible protuberance^ in order to insure the sparks 

 always passing in the same place. One plate was fixed in a horizon- 

 tal position ; the other was suspended, by a thin ivory handle perpen- 

 dicular to its plane, to a glass rod placed in the telescope socket of a 



