92 



TEMPERATURES CALCULATED FROM QUANTITIES OF ELECTRICITY. 



water in the vessel A A is then gradually raised to the boiling point by means of a 

 spirit-lamp, and kept at that temperature until the galvanometer needles and the ther- 

 mometer are quite steady. The same plan must be followed when any other tempera- 

 ture than 212 is under trial, for the thermo-electric wires changing their temperature 

 more rapidly than the mercury in the thermometer, it is absolutely necessary to continue 

 the experiment for some minutes to bring both to the same state of equilibrium. 



340. When a temperature higher than 212 Fah., but under a red heat, is required, 

 I substitute, in place of the vessel A A, a tubulated retort, the tubulure of which is 

 large enough to allow the passage of the bulb of the thermometer and the wires. A 

 quantity of mercury sufficient to fill the retort half full is then introduced, and the tubu- 

 lure being closed by appropriate pieces of soapstone, the neck of the retort is inclined 

 upward, so that the vapour as it rises may condense and drop back again without in- 

 commoding the operator. As in the former case, it is here also necessary to continue 

 each experiment for a few minutes, to bring the thermometer and the thermal pair to 

 the same condition. There is not much difficulty in obtaining any required tempera- 

 ture, by raising or lowering the wick of the lamp. 



341. The metals I have tried were in the form of wires. They were in the state 

 found in commerce, and therefore not pure ; they were obtained in the shops of Phil- 

 adelphia 



TABLE I. 



In this table I have estimated the temperature of boiling mercury at 662 Fahr. 

 The quantities of electricity evolved, as estimated by the torsion of a glass thread, are 

 ranged in columns under their corresponding temperatures. Each series of numbers is 

 the mean of four trials, the differences of which were often imperceptible, and hardly 

 ever amounted to more than one degree. 



342. Now if this table be constructed, the temperatures being ranged along the axis 

 of abscissas, and the quantities of electricity being represented by corresponding ordi- 

 nates, we shall have results similar to those given in fig. 32, in which it is to be ob- 

 served that the curves given by the systems of silver and iron, copper and iron, and pal- 

 ladium and iron, are concave to the axis of abscissas ; but those given by platina and 

 copper, silver and palladium, and platina and iron, are convex. 



343. Let us now apply the numbers obtained by these several pairs for the calcula- 

 tion of temperatures, which will set their action in a more striking point of view. The 

 following table contains such a calculation, on the supposition that for the 90 degrees 

 from 32 Fahr. to 122 Fahr., the increments of electricity are proportional to the tem- 

 peratures. 



