96 



The quantities denoted by and F in the pi'eceding equation 

 being both positive, it is concluded that, when a thermo-electric current 

 passes through a piece of iron from one end kept at about 240 

 cent., to the other end kept cold, in a circuit of ivhich the remainder 

 is copper, including a long resistance wire of uniform temperature 

 throughout or an electro-magnetic engine raising weights, there is 

 heat evolved at the cold junction of the copper and iron, and (jio heat 

 being either absorbed or evolved at the hot junction^ there must be a 

 quantity/ of heat absorbed on the whole in the rest of the circuit. 

 When there is no engine raising weights, in the circuit, the sum of 

 the quantities evolved, at the cold junction, and generated in the 

 " resistance wire^'' is equal to the quantity absorbed on the whole 

 in the other parts of the circuit. When there is an engine in the 

 circuit, the sum of the heat evolved at the cold junction and the 

 thermal equivalent of the weights raised, is equal to the quantity of 

 heat absorbed on the whole in all the circuit except the cold junction. 

 7. An application of the theory to the case of a circuit consisting 

 of several different metals, shews that if 



p(A,B), ^(B,C), <p(C,J)), . . . <p(Z,A) 

 denote the electromotive forces in single elements, consisting respec- 

 tively of different metals taken in order, with the same absolute 

 temperatures of the junctions in each element, we have 



9 (A, B) + ^ (B, C) + ?) (C, D) . . . +p(Z,A) = 0, 

 which expresses a proposition, the truth of which was first pointed 

 out and experimentally verified by Becquerel. A curious experi- 

 mental verification of this proposition (so far as regards the signs of 

 the terms of the preceding equation) was made by the author, with 

 reference to certain specimens of platinum wire, and iron and copper 

 wires. He had observed that the platinum wire, with iron wires 

 bent round its ends, constituted a less powerful thermo-electric ele- 

 ment than an iron wire with copper wires bent round its ends, for 

 temperatures within atmospheric limits. He tried, in consequence, 

 the platinum wire with copper wires bent round its ends, and con- 

 nected with the ends of a galvanometer coil ; and he found that, 

 with temperatures within atmospheric limits, a current passed from 

 the copper to the platinum through the hot junction, and concluded 

 that, in the thermo-electric series 



Antimony, Iron, ( p^L, } bismuth, 



