﻿On the Electromotive Force on the Contact of different Metals. 19 



the ends, opposite to each other, and at right angles to the faces. 

 Through these tubes passed the metal wire to be investigated, 

 consisting of two wires soldered together, in such a manner that 

 the place of soldering was just in the middle of the cylinder. 

 On the tubes in question caoutchouc tubes were fitted, which, to 

 make them air-tight, were firmly tied both over the copper tube 

 and over the metal rods. With the same object the space between 

 the wire and the metal tube was filled with melted wax. To 

 preserve the wire from conducting communication with the cop- 

 per cylinder, the two tubes were lined internally with caoutchouc. 

 At one end of the copper cylinder a metal tube, bent at right 

 angles and provided with a stopcock, was soldered, to which was 

 fitted air-tight a glass tube. This tube was provided with a scale 

 divided into millimetres. The copper cylinder with the glass tube 

 was fastened on a mahogany board, which could be placed, and 

 secured by screws, at different angles with the horizontal. In 

 all the experiments the glass tube formed an angle of 5° with 

 the plane in question. The copper cylinder had such a posi- 

 tion that the wire traversing it was horizontal. After the wire 

 to be investigated had been introduced, the cylinder was tested 

 by fixing a small compression-pump to the metal tube and ob- 

 serving the pressure on a water-manometer in connexion with 

 the tube. To keep the temperature around the copper cylinder 

 constant, it was surrounded by a double cylindrical zinc jacket 

 which held 5*85 litres of water. The internal diameter of the zinc 

 jacket was 150 millims., so that the annular space between this 

 and the copper cylinder was 35 millims. wide. The zinc jacket 

 was in four parts, so that it could be conveniently inserted in its 

 position ; and suitable apertures were provided for the passage of 

 the wire and of the above-mentioned metal tube. The zinc 

 jacket at the time of the experiments, was filled with water which 

 had the same temperature as the air in the room. In this man- 

 ner the zinc jacket acquired a temperature which did not mate- 

 rially vary during the short time which was necessary for an 

 experiment. On the further end of the glass tube a short caout- 

 chouc tube was placed, which was connected with an open cylin- 

 drical metal box 55 millims. in diameter. Into this box the 

 liquid was poured which was to serve as index in reading off. By 

 raising or lowering the box the column of liquid could be made 

 to stand at any suitable position in the tube ; the internal dia- 

 meter of this tube was about 2'5 millims. When the tube was 

 too narrow, the capillarity acted so strongly that the column of 

 liquid was seriously impeded in its motion, and the air-thermo- 

 meter thereby became less delicate. In the first four of the sub- 

 sequent experiments spirit was used as index, and in all the 

 others a coloured liquid containing spirit (cognac) . 



C2 



