of Electricity through Rarefied Air. 7 



galvanometer and the bridge. All the series of observations 

 agreed in this: — that if the conduction between k and the 

 rarefied-air space was interrupted, and consequently the entire 

 discharge passed to q and was there divided between the 

 bridge and the galvanometer to arrive at the ball /, the deflec- 

 tion did not amount to more than 1 or 1'5 division of the 

 scale; but if the discharge divided at k and part of it passed 

 through the rarefied air, the deflection might amount to 50 

 scale-divisions — although a lower figure might have been ex- 

 pected, seeing that in this case a smaller portion of the dis- 

 charge passed through the galvanometer. 



The question now is, to what cause must we ascribe the cir- 

 cumstance that the deflections were several times greater in 

 one case than in the other ? Perhaps the first thought that 

 will occur to us is that the discharge is oscillatory. One 

 might then say: — "At the first swing positive electricity is 

 discharged, without dividing at k, through the bridge n and 

 the galvanometer to arrive at the ball/; when, at the second 

 swing (a little weaker than the first), the electricity returns, 

 it takes its path to h exclusively through the rarefied-air space; 

 at the third swing it again passes exclusively through the gal- 

 vanometer and the bridge, while at the fourth it is discharged 

 only through the rarefied-air space, and so on. In this way 

 the deflections would be greater when k is in metallic commu- 

 nication with the rarefied air than when the communication 

 between them is interrupted." But it will readily be perceived 

 that an oscillatory discharge of this kind is absolutely impos- 

 sible. The same direction and the same amount of deflection 

 have been observed, whether the point of interruption between 

 the balls e and / was near to & t , as shown in the figure, or 

 placed near to %. It can therefore be said that every thing 

 is symmetric around the rarefied-air space, and hence there is 

 no valid reason for the electricity, in passing from left to 

 right, to be brought to choose another path than that which 

 it takes when it returns from right to left. Further on, more- 

 over, will be found some proofs demonstrating that the oscil- 

 latory discharge cannot be the cause of the great deflections 

 mentioned above. 



Others will perhaps be disposed to assume that the reason 

 of the great deflections which take place when the point k is 

 connected with the rarefied-air space, is the rise of induction- 

 currents, at the time of the discharge, in the helix of the gal- 

 vanometer or at some other point of the closed circuit, and 

 that to these currents we must attribute the above-mentioned 

 deflections. Now, further on will likewise be given proofs 

 that the great deflections have just as little to do with induced 



