978 
electromotive equilibrium with copper and silver, contains but 
exceedingly little silver. 
As was shown before‘) there then exists between the copper and the 
silver wire still a potential difference equal to the Volta-effect. Now 
the tube 6 was placed in a thermostat a and the two soldering 
places Ag—Cull and Cu l—Cu II in melting ice. After the thermostat 
had been kept constant at a definite temperature for some time, the 
electromotive force of the circuit was measured. This measurement, 
ranging over a temperature interval of from 40°—80°, gave no 
indications of a discontinuity in the neighbourhood of 60°, after 
which it was resolved to apply a more accurate measurement, 
which we found in a method, which we shall call the diferential 
method. The arrangement of this differential method is given in 
outline below in Fig. 2. 
Cull Cul As Cul Cul 
a 
Fig. 2. 
The two extremities of the silver wire, which are soldered to 
wires of pure copper, are now both in the thermostat a, but the 
capillary & is filled with a solution of copper sulphate, and c 
contains anhydrous paraffin oil. Both solutions are under a nitrogen 
atmosphere *). 
The consideration that led to this arrangement, is as follows. So 
long as no transition point is reached on change of temperature, 
the electromotive force of the circuit will be very small, hence it 
will change but exceedingly little with the temperature in the ther- 
mostat; but when a transition temperature of silver or copper is 
passed, it is probable that the thread of the differential thermoele- 
ment, which is in electromotive equilibrium with the electrolyte, is 
1) These Proc. Vol. XX1, 3, p. 386. 
2) In reality the tubes b and c, which contain the metal wires immersed in a 
liquid, have been first exhausted of air, then filled with nitrogen, and finally 
fused to. 
