330 THE ELECTRO-MOTIVE POWER OF HEAT. [MKMOIR XXIV. 



the increments of the quantities of electricity are propor- 

 tional to the increments of temperature throughout the 

 range of the mercurial thermometer. 



Let us now pass to the second proposition : "That the 

 tension undergoes a slight increase with increase of tem- 

 perature a phenomenon due to the increased resistance 

 to conduction of metals when their temperature rises." 



It will be seen, on consulting the following table, that 

 pairs of different metals at the same temperature have 

 tensions which are apparently very different. 



The currents the tensions ot which are here indicated 

 were generated by keeping one end of the thermal pair 

 in boiling water, the other ends being maintained at a 

 temperature of 32 Fahr. 



TABLE III. 



We perceive, therefore, that there apparently exist 

 specific differences in the qualities of electric currents 

 derived from different sources. If, for example, we take 

 a pair of platinum and palladium and expose it to a tem- 

 perature which shall generate a current capable of de- 

 flecting the torsion balance through 1000, and then 

 obstruct it by a wire of such dimensions as to stop one 

 half, or only allow 500 to pass, and repeat the experi- 

 ment with a current generated by bismuth and antimo- 

 ny, the temperature being still so adjusted as to give a 

 deflection of 1000, on making this pass through the 

 same intercepting wire, perhaps not much more than one 

 eighth of it will go through the galvanometer. 



This peculiarity of thermo-electric currents depends 



