hy Thermo-electric Currents. 413 



extreme mobility of the needles renders the manipulation 

 very difficult. 



The difficulties which are met with in the accurate mea- 

 surement of the intensities of thermo-electric currents by the 

 galvanometer and sine compass, induced me to seek for a 

 process of measurement which was wholly independent of 

 those instruments ; and I think I have succeeded by the fol- 

 lowing method, which appears to me susceptible of successful 

 application to the study of the laws of thermo-electric currents. 



I constructed an element of bismuth and antimony com- 

 posed of two bars, ABCD, Plate 1. figs. 1 and 2, cast in the 

 same mould. These two perfectly similar bars are juxtaposed 

 throughout their extent and kept separate by a blade of ivory ; 

 they only touch at the two extremities, A and D, where the 

 two solderings are. The length BC is 20 centimetres, the 

 vertical pieces ABCD are 12 centimetres. This element of 

 bismuth and antimony is the normal element, with which I 

 compare all the other thermo-electric elements, but it should 

 not be used for temperatures exceeding 30°. 



The element destined for high temperatures is formed of 

 an iron wire and a platinum wire of 1 millimetre in diameter; 

 the extremities of these wires are soldered with silver. The 

 iron wire, E/F, fig. 4, is about 80 centimetres in length ; the 

 two platinum wires, Ec, F</, are fixed near to the iron wire, 

 from which they are isolated by a non-conducting envelope. 

 In their lower portion these wires are separated by a plate of 

 thin glass. They are terminated by two brass appendages, c 

 and d, which allow of the introduction of a galvanometric ap- 

 paratus into the circuit. 



The two junctions, E and F, are kept in glass tubes filled 

 with a fixed oil containing no oxygen. One of these tubes is 

 placed in the oil-bath mentioned in a former part of this work 

 (p. 57), by the side of a mercurial thermometer extending from 

 0° to 350°. In some experiments an air-thermometer placed 

 in the bath was employed. The tube which contained the 

 second junction is kept at a constant temperature by means 

 of melting ice, or placed in a great water-bath alongside the 

 mercurial thermometer. 



The normal element of bismuth and antimony, ABCD, 

 figs. 5 and 6, is arranged in such a manner that the two sol- 

 derings, A and D, dip into two vessels, MN, M'N', filled with 

 water at different temperatures, and separated from each other 

 by a partition, SR. The same agitator, FGF'G', serves to 

 agitate at the same time the water in both vessels, and two 

 very accurate and strictly comparable thermometers, T and 

 T', are placed near the two solderings. These thermometers 



