332 THE ELECTRO-MOTIVE POWER OF HEAT. [MEMOIR XXIV. 

 As 783 : 163 :: 123 : 25| instead of 23, 



showing, therefore, a slight rise of tension. 



2d. The pair of copper and ifcon gave at the boiling- 

 point of water 300, of which 57 passed the secondary 

 wire. The temperature was now raised, with the follow- 

 ing results : 



400 degrees passing the primary, 95 the secondary wire. 



553 " " " ' 113 " 



545 " " " 112 



493 " " " 110 " " 



It will be understood that although the quantities of 

 electricity indicated in the first column do not regularly 

 increase, the temperatures were, notwithstanding, going 

 regularly upwards: to this peculiarity of the systems 

 into which iron enters I have already alluded. Let us 

 now compare these measures with those obtained for the 

 boiling-point of water : 



As 490 : 95 

 553 : 113 

 545 : 1 1 2 

 493 : 1 10 



300 : 58 instead of 57. 

 300 : 01 " 

 300 : 61 " 

 300 : 67 " 



We find, therefore, that in the case of both these systems 

 of metals the tension slowly rises with increase of tem- 

 perature, being much better marked in the latter than in 

 the former instance. 



The increase of tension here detected depends unques- 

 tionably on increased resistance to conduction, which the 

 wires exhibit as their temperature rises, as the following 

 experiments show. 



A pair of copper and iron evolved a current at the 

 boiling-point of water, which, passing through a wire of 

 copper eight feet long, was determined at the galvanom- 

 eter to be 176. Having twisted a part of this wire into 

 a spiral, so as to go over the flame of a spirit-lamp, eight 

 inches of it were thereby brought to a red heat; the de- 



