48 



PROCEEDINGS OF THE AMERICAN ACADEMY 



It is, therefore, of the greatest importance to use absolutely pure 

 substances in all experiments on thermoelectricity. In the following 

 experiments, therefore, the metals used have always been chemically 

 pure, — deposited by electrolysis. 



The first experiments were made with copper and nickel. The 

 nickel was about twelve inches long and V-shaped. To each end was 

 soldered a strip of copper, about eight inches long. During the ex- 

 periment one end was placed in melting snow and the other in a 

 mixture of snow and calcic chloride, — 30 grammes of snow to 40 of 

 calcic chloride. The two junctions, and also the metals as far as they 

 were in contact with the freezing mixture, were coated with shellac. 

 A mercury thermometer — the zero point of which had previously 

 been verified — was bound to the colder junction of the copper and 

 nickel. Copper wires, soldered to the free ends of the two strips of 

 electrolytic copper, connected the thermoelectric element with the gal- 

 vanometer, the circuit being made or broken by means of a key. 

 The galvanometer was a Thomson's mirror galvanometer of low re- 

 sistance. The mixture of snow and CaCl2 generally gave a tempera- 

 ture of about — 25° C. From this the temperature gradually 

 rose, and at every 5° increase the circuit was made, and the deflec- 

 tion of the galvanometer observed. When the temperature had 

 reached 0° the junction was placed in a vessel of water and heated 

 gradually to about 80° C, the deflection of the galvanometer being ob- 

 served for every ten degrees increase of temperature. The following 

 are the results from five series of experiments. The junctions were 

 reversed in every alternate series ; that is, the junction that, in the 

 first, was placed in melting snow, was, in the second, placed in CaClj 

 and snow. This obviated whatever irregularities might arise from 

 any difference between the two junctions. 



The fii'st column gives the temperature of the colder junction. The 

 second the number of the experiment. The third the resistance of 

 the whole circuit and the galvanometer. The resistance of the latter 

 was six ohms. The fourth gives the deflection of the galvanometer 

 needle. The fifth gives the product of this deflection into the resist- 

 ance, which is directly proportional to the electromotive force. 



