58 J. A. EWING AND J. G. MACGREGOR ON THE 
therefore occurred to us that the Wheatstone bridge test would be applicable 
provided we could observe the znstantaneous effect of a current upon the solu- 
tion. This we accomplished in the following manner :— 
The solutions were placed in a glass tube, narrow along the central part, 
and wide at the ends (see figure). The narrow part was about 15 centimetres 

in length, and 0°35 square centimetres in internal cross section, The current 
passed between platinum plates, connected with platinum wires passing through 
the bungs which closed the tube. The ratio of the cross section of the wide 
parts to that of the narrow part was so great as to permit small variations in the 
positions of the plates, without perceptibly changing the electrical resistance of 
the solution. The resistance coils used were in the form of a Wheatstone 
bridge, as arranged by Exuiott Brothers, London. The galvanometer was an 
ordinary “ dead beat” mirror one, of small resistance. It is essential that in 
this test the inertia of the mirror and needle should be very small. The electro- 
motive force was that of one Grove’s cell. The electrodes of the tube dipped 
into mercury pools a and a. The “rockers” were arranged together on one 
board. By means of the commutator f a could be connected with 0 or 0’, and 
a’ with 6’ or b. In this way the alternate currents could be sent through the 
tube in opposite directions ; this tended to diminish polarisation. The “‘ battery 
circuit” was completed when d and d’ were connected,—the “ galvanometer 
circuit” when ¢ and é’ were connected ; ¢ and ¢c’ were kept in permanent con- 
nection with the electrodes of the tube by means of the fixed wires ac and a’c’. 
The battery circuit and galvanometer circuit were simultaneously closed by 
means of the rockerg. This consisted of a 4 frame of insulating material, with a 
cross wire at each of the three extremities, bent like an inverted U. The six 
