152 Prof. J. J. Thomson on the 



We will begin by considering how metallic conduction 

 differs from electrolytic. In the first place, as the tempera- 

 ture increases the conductivity of electrolytes, as a general 

 rule, increases, while that of metals diminishes. This rule 

 is not, however, without exceptions : there are cases in which, 

 though the conduction is not usually supposed to be electro- 

 lytic, the conductivity increases as the temperature increases. 

 Carbon is a striking instance of this, and quite lately Feussner 

 has prepared alloys of manganese and copper whose con- 

 ductivities show the same peculiarity. These exceptions are 

 sufficient to show that increase of conductivity with the tem- 

 perature is not a sufficient test to separate electrolytic from 

 metallic conduction. 



If we regard the passage of electricity through a body as 

 essentially bound up with chemical changes, it does not seem 

 surprising that an increase in temperature may produce 

 opposite effects on the conductivities of two substances, even 

 though in both cases the conduction was effected by changes 

 in the pairings of the atoms. For the action of an increase 

 of temperature has a two-fold effect on the processes which, 

 on this view, accompany electric conduction. In the first 

 place, it may promote the splitting up of the molecules into 

 atoms which, on this theory, forms one part of the process of 

 conduction ; but, on the other hand, after the molecules are 

 split up it retards their reunion, which forms another part of 

 the process. And, again, an increase in the temperature 

 increases the distance between the molecules, and this will 

 also retard the rate at which chemical interchange takes 

 place. The fact that the metals are solids is no reason why 

 the conductivity through them should not be electrolytic in 

 its nature, for there are many instances of solid electrolytes ; 

 thus Lehmann has shown that electrolysis takes place through 

 a crystal of silver iodide placed between silver electrodes 

 without any change being perceptible in the shape or size of 

 the crystal, though it was watched through a microscope 

 whilst the current was passing. 



With regard to the appearance of the products of chemical 

 decomposition at the electrodes, we could not expect to get 

 any evidence of this in the case of the elementary metals ; the 

 case of alloys seems more hopeful ; but Professor Roberts 

 Austen has examined several alloys through which a powerful 

 current had been passed without detecting any difference in 

 the composition of the alloy at the terminals. This result 

 does not, however, seem to me to prove that the conduction 

 was not electrolytic; for some alloys are little more than 

 mixtures, whilst others behave as if they were solutions of one 



