1886.] 



A Theory of Voltaic Action. 



297 



so that after raising the rod C it could be readily brought back to its 

 original position. The quadrants were now levelled by a spirit-level 

 from front to back, after which the difference of potential might be 

 at once taken, or they might be removed again to be cleaned, rapidly 

 replaced, and a reading taken as quickly as possible afterwards with- 

 out the loss of time in adjusting them. A " reading " was taken by 

 causing the needle (by reversing its potential twice) to deflect first on 

 one side (say right) of zero, then to the left, and again to the right. 

 The mean of the two right-hand deflections was added to the left-hand 

 one in order to eliminate the effect of a slow shifting of the zero 

 which sometimes occurred, caused apparently by a slow untwisting 

 of the suspension wire, or a slight warping of the vulcanite piece M. 



7. The readings were either qualitative when the two quadrants 

 were put in metallic connexion, or quantitative when the difference of 

 potential of their films was compared with a standard cell. 



In order to make the connexions conveniently for both kinds of 

 readings a modified Pohl's commutator was used, by means of which 

 the quadrants could either be directly connected with each other, or 

 else one to each pole of the standard cell W in either sense. 



8. At first two standard cells were used, a Clark's cell purchased 

 from Elliot Brothers, and a Daniell constructed as at W. X, Y, Z 

 are three test-tubes fixed in a block of paraffin. X contains a half- 

 saturated solution of zinc sulphate, with a rod of redistilled zinc not 

 amalgamated. In Z is a gutta-percha covered copper wire, having 

 its lower end cleaned, wound into a spiral, and covered with crystals 

 of copper sulphate and water. Y contains distilled water with a strip 

 of zinc to reduce any copper salt that might diffuse over from Z. 



When the Clark's cell was used, the results were reduced to the 

 Daniell as unit value. Comparisons with both cells were at first 

 frequently made. 



9. The measurement of the difference of potential near any pair of 

 quadrants was made by a well-known method as follows. The cell 

 was first connected, so that its electromotive force should increase the 

 deflection due to the difference of potential near the quadrants, and a 

 reading a taken. The connexion was then reversed, and a reading b 

 taken. If d is the deflection due to the standard cell, and p that due 

 to the difference of potential near the quadrants — 



d-\-p = a 

 d—p = b 



therefore ^ = - — ~. 



d a + b 



For example, in comparing the difference of potential of the films on 

 slightly tarnished copper and zinc in air the following deflections were 

 found — 



x 2 



