."'7i' Thermo-electric Properties of Salt Solutions. [Apr. It 



whence H = O1956, 



H/* = 6-815 x 10- 4 . 



Current 335 by the second method gave perfectly constant cl 

 34 of resistance for both directions of current, 



whence 



H = 0-2078, 

 H/< = 7-24 x 10- 



Finally, the measurement was reduced to a null method by 

 two pieces of apparatus like A, of the same size, and as nearly ali 

 as possible. These were used side by side in a wooden bowl 

 liquid of an oblong shape, with a copper bottom, and divided by 

 ebonite plate, so that the current from the electrodes should go 

 tically through the liquid. Each platinoid coil had a resistance 

 12'35 ohms., and the thermometer coils were 32 and 36 ohms, 

 ively. The total thickness of each, bobbin with its electrode 

 was not much above % in., the diameter a little over 1 in. 



The current passes through one of the 7 coils, and from one eU 

 ode plate .to the other through the bottom of the liquid. It 

 adjusted so that the ratio of the resistances of the thermometer 

 should remain constant for both directions of current. 



We now have, using the same letters as before, the single eqt 

 tion 



AC 2 +HC = AC 2 -HC-rRC 2 , 



assuming everything symmetrical, 



or H = RC. 



If there is any want of symmetry, we shall get two values for I 

 one for each direction of the current, and if these are nearly equ 

 we are safe in taking their mean. 



With this apparatus for the same solution a perfect balance 

 got in each direction for the same value of C = 0'0339 amp. 



.-. H = 6-165. C = 0-2091 volt amp. 

 Hjt = 7-26 x 10- 4 . 



As this seemed rather high, the contact of the electrodes with 

 bobbins was made better by amalgamating their surfaces. 



A current 280 potentiometer gave a bridge change 10 in 5'. 

 260 +17 



Mean 272'6 should give balance. 



Current 272 gave perfect balance in each direction for diffe 

 lengths of time. / 



