97 



this .platinum wire must, at ordinary temperatures, be between iron 

 and copper. He found that the platinum wire retained the same 

 properties after having been heated to redness in a spirit-lamp and 

 cooled again ; but with temperatures above some limit itself consider- 

 ably below that of boiling water, he found that the iron and platinum 

 constituted a more powerful thermo-electric element than the iron 

 and copper ; and he verified that for such temperatures, in the pla- 

 tinum and copper element the current was from the platinum to 

 the copper through the hot junction, and therefore that the copper 

 now lay between the iron and the platinum of the series, or in the 

 position in which other observers have generally found copper to lie 

 with reference to platinum. A second somewhat thinner platinum 

 wire was found to lie invariably on the negative side of copper, for 

 all temperatures above the freezing point ; but a third, still thinner, 

 possessed the same property as the first, although in a less marked 

 degree, as the superior limit of the range of temperatures for which 

 it was positive towards copper was lower than in the case of the first 

 wire. By making an element of the first and third platinum wire, 

 it was found that the former was positive towards the latter, as was 

 to be expected. 



In conclusion, various objects of experimental research regarding 

 thermo-electric forces and currents are pointed out, and methods of 

 experimenting are suggested. It is pointed out that, failing direct 

 data, the absolute value of the electromotive force in an element of 

 copper and bismuth, with its two junctions kept at the temperatures 

 0° and 100° cent., may be estimated indirectly from Pouillet's com- 

 parison of the strength of the current it sends through a copper 

 wire 20 metres long and 1 milHmetre in diameter, with the strength 

 of a current decomposing water at an observed rate ; by means of 

 determinations by Weber, and of others, of the specific resistance of 

 copper and the electro-chemical equivalent of water, in absolute units. 

 The specific resistances of different specimens of copper having been 

 found to differ considerably from one another, it is impossible, with- 

 out experiments on the individual wire used by M. Pouillet, to deter- 

 mine with much accuracy the absolute resistance of his circuit, but 

 the author has estimated it on the hypothesis that the specific resist- 

 ance of its substance is 2i British units. Taking '02 as the electro- 

 chemical equivalent of water in British absolute units, the author has 

 thus found 16300 as the electromotive force of an element of cop- 

 per and bismuth, with the two junctions at 0°and lOO'' respectively. 



