'ELECTRICITY. 95 



upon the view I am suggesting, the discharge is itself this 

 affection of matter : and the writing these passages affords, 

 to me at least, a striking instance of how much ideas are 

 bound up in words, when, to express a view differing from 

 the received one, words involving the received one are neces- 

 sarily used. 



Passing now to the effect of the transmission of electri- 

 city by the class of the best conducting bodies, such as the 

 metals and carbon, here, though we cannot at present give the 

 exact character of the motion impressed upon the particles, 

 there are yet many experiments which show that a change 

 takes place in such substances when they are affected by elec- 

 tricity. 



Let discharges from a Leyden jar or battery be passed 

 through a platinum wire, too thick to be fused by the dis- 

 charges, and free from constraint, it will be found that the 

 wire is shortened ; it has undergone a molecular change, and 

 apparently been acted on by a force tranverse to its length. 

 If the discharges be continued, it gradually gathers up in 

 small irregular bends or convolutions. So with voltaic elec- 

 tricity : place a platinum wire in a trough of porcelain, so 

 that when fused it shall retain its position as a wire, and then 

 ignite it by a voltaic battery. As it reaches the point of fu- 

 sion it will snap asunder, showing a contraction in length, and 

 consequently a distension or increase in its transverse dimen- 

 sions. Perform the same experiment with a lead wire, 

 which can be more readily kept in a state of fusion, and fol- 

 low it, as it contracts, by the terminal wires of the battery ; 

 it will be seen to gather up in nodules, which press on each 

 other like a string of beads of a soft material which have 

 been longitudinally compressed. 



As we increase the thickness of the wires in these exper- 

 iments with reference to the electrical force employed, we les- 

 sen the perceptible effect : but even in this case we shall be 

 enabled safely to infer that some molecular change accompa- 



