﻿on the Contact of different Metals. 21 



be s, and m be a constant, this quantity of heat will be expressed 

 by ms 2 . On the other hand, according to the direction of the 

 current, an amount of heat is produced or absorbed at the con- 

 tact which is proportional to the intensity of the current, and 

 may be designated by +ns, in which n is another constant. 

 We thus find ms 2 ±ns = A-\-a. If now the direction is reversed, 

 the index after a few minutes is stationary in a new position a', 

 and we get ms 2 + ns = A + a' } in which the sign before the second 

 term is altered. From this we obtain +ns = a — a'. The dif- 

 ference between the two stationary positions of the index when 

 the current first goes in one and then in the other direction thus 

 becomes a measure for the quantities of heat in question. 



The above method of obervation would furnish very rigorous 

 determinations of the quantities of heat in question if the heat- 

 ing of the air in the copper cylinder were the only cause of the 

 motion of the column of liquid. But this is far from being the case . 

 Variations in the barometric height during even the few minutes 

 which are required to effect two successive readings are frequently 

 sufficient materially to alter the position of the column of liquid. 

 One day, on which a violent storm prevailed, no observations could 

 be made; for, owing to variations in the pressure of the atmosphere, 

 the column of liquid was continually moving backwards and for- 

 wards. This, however, is exceptional. To avoid this unplea- 

 santness as far as possible, the experiments were made thus : — 

 After the current had circulated for a sufficient length of time, 

 the position of the column of liquid was observed ; the current 

 was then reversed and the new position of equilibrium observed, 

 upon which the current was again reversed and the first observa- 

 tion repeated. Of the first and third readings the mean was taken; 

 and the same course was continued until so many determinations 

 had been obtained that the mean was sufficiently accurate. As 

 the time between two successive observations was approximately 

 equal, accurate results must thus have been obtained, assuming 

 that the barometric pressure varied uniformly. It would not 

 have been impossible to procure accurate corrections for the 

 errors which alterations in the barometric height produce. If, 

 for instance, we had two perfectly equal air-thermometers, one of 

 which, as is here the case, was used exclusively for determining 

 the voltaic heating, and the other for measuring variations in 

 the pressure, by simultaneously reading the two the necessary 

 corrections would be obtained. The same object would also be 

 obtained if the glass tube were not allowed to communicate with 

 the air through the open metallic box, but with a large air-tight 

 closed vessel whose temperature was kept quite constant. Mean- 

 while the method used was considered adequate to the object in 

 view. We pass now to the proper observations. 



