472 Mr. K. Colley on a Case of Work 



This force is independent of the intensity of the current pass- 

 ing through the liquid, and proportional to the height of the 

 column. 



The principal difficulty of the experiments was the presence 

 of inconstant and casual currents, due to the inequality of the 

 electrodes, their polarization, <fcc. The intensity of those cur- 

 rents far exceeded that of the current given by the electro- 

 motive force e, if special precautions were not taken to weaken 

 them. I found them much enfeebled by removing the air and 

 gases dissolved in the liquid. 



After some trials, I fixed upon the following order of appa- 

 ratus : — The liquid column was a glass tube filled with the 

 solution of the electrolyte, closed at one extremity, and com- 

 municating at the other with an air-pump, through a 

 caoutchouc tube, which could be closed by a pinchcock 

 close to the glass tube. The electrodes were sealed in two 

 lateral tubules, near the extremities. The tube turned on a 

 horizontal axis passing through its centre of gravity, so that 

 either of its ends could be directed upwards. Thus, without 

 any change in the rest of the apparatus, the current could be 

 made f) ascend or descend in the liquid column. One of the 

 tubes which I employed, and Avhich I will call No. 1, had a 

 length of 1*6 metre. Filled with a solution of nitrate of 

 silver, it offered a resistance of 774 Siemens's units. Tube 

 No. 2 had a length of 3*6 metre, and a resistance nearly double 

 that of No. 1. 



The galvanometer was a reflecting one and very sensitive. 

 The wire of the induction-coil was very fine and made a great 

 number of turns; its galvanic resistance amounted to 4077 

 Siemens's units. Either of the two tubes together with the 

 galvanometer forming a closed circuit, and the sensitiveness 

 of the latter instrument having been determined by prelimi- 

 nary experiments, the deviation due to the electromotive 

 force e acting in a circuit of which the resistance r was 

 known could be calculated beforehand. As I always observed 



E -f- € E — 6 6 



the difference =2— of the descending and the 



r r r 



ascending current, the difference of the deviations produced 

 by these two currents, which we will name 8, was the double 

 of the preceding ; and as by " deviation " I mean everywhere 

 the difference of the elongations on both sides of the position 

 of the equilibrium, elongations obtained on changing by a 

 commutator the direction of the current in the galvanometer- 

 coil alone ; it is evident that 8 was four times a simple devia- 



tion due to the current — . The value of 8 is expressed in 



