200 Profs. W. E. Ayrton and John Perry. [Mar. 21, 



TT, which support the wooden frame WPW, and the positions of which 

 on the three wooden blocks T are fixed by " hole, slot, and plane." One 

 of these levelling screws has been removed in the drawing to show the 

 lifting arrangement. 



F and G are strips of brass soldered to the plates A and B, and 

 connected by means of clips to wires leading to the insulated electrodes 

 of the quadrant electrometer, and to the insulated screws A', B', of the 

 short-circuit key. 



In the drawing we have chosen to arrange the apparatus for deter- 

 mining the electromotive force of contact between a metal D and a 

 liquid 0. VS is a screw-jack which enables the oblong wooden table 

 OS' 0' to be raised to any level, and the table may be turned round 

 horizontally at any level. Three levelling screws, working in nuts in 

 the table, support the metal plate D, and three others, exactly similar, 

 support the porcelain vessel G which contains the liquid, so that the 

 upper surfaces of both G and D are in the same horizontal plane. A 

 binding screw a is soldered to the plate ; in the present case it holds 

 the metal strip E : which makes contact with the liquid ; E must have 

 been cut from the same sheet as the plate D. When there are two 



face C, and a + v the potential of D at the beginning of the experiment. When A 

 and B are connected to earth, the charge on A and its electrometer-quadrants is 



4<7rt 



Y being the potential of the needle. The charge on B and its quadrants is 



After reversal, if C has a potential y and Z>, therefore a potential y + v, the charge on 

 A and its quadrants is 



And the charge on B and its quadrants is 



-^fr^ + { s+ 5M < Ya - Vb ) } < Vb - y ) 



where Ya and Yb are the potentials of A and B respectively. Equating the re- 

 spective charges before and after reversal, and subtracting, we get 



< Ta - Tb) (^ + 0- ,b(Va2 - Tb2)=2 £ ■ • • (1) 



If we might neglect the change of capacity of the quadrants, we should have 

 (Y A -Y B /-^ + 5 \ 



tea \47rt J 



2a 

 4^t 



• (2) 



so that v varies as (Ya —Yb). 



Now our experiments show this to be the case, and consequently (2) gives the 

 correct value of v. 



