XXIV. J TIIK KLECTKO-MOTIVK I'OWKlt OF HEAT. 



on the conducting resistance of the system that generates 

 them. It is possible to give a current a higher or a 

 lower tension by simply making use of thin or thick 

 wires to generate or to carry it. In the foregoing table 

 the current from platinum and palladium had a high 

 tension, because slender wires of those metals happened 

 to be used to generate it ; and the current from antimo- 

 ny and bismuth had a low tension because thick bars 

 of those substances were employed. In the former case, 

 the conducting resistance was greater than in the latter, 

 and hence the tension of the current was higher. 



That this is strictly true will appear on examining the 

 current evolved by any number of systems under the 

 same condition of resistance to conduction. 



In a great number of trials I failed in getting any 

 trustworthy results as respects tension of currents at 

 high temperatures, on account of the difficulty of main- 

 taining the thermo-electric pair at the same degree with- 

 out variation. By employing, however, a small black- 

 lead furnace, to which was adapted a covered sand-bath 

 into which the wires could be plunged, I succeeded at 

 last ; for with this arrangement a regulated temperature 

 could be kept up for a length of time. 



The experiment was made with care in the case of 

 two systems of metals: 1st, copper and platinum; 2d, 

 copper and iron. 



1st. At the boiling-point of water, a pair of copper and 

 platinum, the unexcited extremity of which was carefully 

 maintained at 67 Fahr., evolved as a mean of four trials, 

 three of which were absolutely identical, 123 of electric- 

 ity, of which 23 could pass a secondary wire. 



Then, by the aid of the furnace and sand-bath, the 

 temperature was raised until the pair evolved 783 as a 

 mean of four trials; of these 163 could pass the second- 

 ary wire. Now, 



