354 Messrs. Threlfall and Pollock on the Clark Cell 



current is flowing, measured from the value of P.D. when the 

 cell is at rest. Then 



A = P.D. when cell is at rest — P.D. when cell is giving 

 current C, 

 = E-(E-CR), 



= CR when there is no polarization. 

 Now at any instant let the E.M.F. change by polarization so 

 that it becomes E', and let the internal resistance become R". 

 Then, on short-circuiting, 



A = E-(E'-C / R"), 

 _ F E'R" 

 R' + R"" 

 Consequently, if E' diminishes or R" increases, we shall have 

 increased values of A. These increased values will then indi- 

 cate what is generally known as polarization. It is possible, 

 however, that, there may be another sort of polarization (if we 

 may venture to so extend the term) of which an increasing E' 

 or diminishing R" is characteristic. If such an action exists 

 and tends to increase as time goes on, we shall observe a 

 continual decrease of A as time progresses after short- 

 circuiting. 



Such an effect has been pretty constantly observed ; and as 

 it was not due to any change in the compensator, we are 

 obliged to attribute it to a small variation in the Clark cell 

 itself in one or both of the directions mentioned. 



On referring to Wiedemann (Die Lehre von der Electricitat, 

 edition 1883, vol. ii. p. 708) we found a statement, on the 

 authority of Hockin and Taylor (Journ. Soc. Teleg. Eng. viii. 

 p. 282, to which we have not access), to the effect that an 

 increase of thickness of a polarizing layer of hydrogen on mer- 

 cury causes the polarized surface to become more positive. 

 The method employed, and the statement that finally bubbles 

 of gas may be given off by sufficiently diminishing the area of 

 the mercury surface, suggests that the polarizations used in 

 the experiments were very strong. The electrolyte was dilute 

 sulphuric acid. 



Some experiments were made by us, using a similar but 

 improved method and with zinc-sulphate solution as an elec- 

 trolyte. The sort of effect we were in search of was an 

 increase of potential on one mercury surface of about 0*0002 

 volt. Every precaution was taken in securing mercury and 

 sulphate of zinc identical with those used in Clark cells, and in 

 checking the behaviour of the electrodes. An effect larger 

 than that we were in search of could generally be detected on 



