RESISTANCE OF ELECTROLYTES. 28 
That the currents thus produced should be capable of being substituted for the mag- 
neto-electric currents, it was necessary not only that the commutators should be accurately 
made, but also that the rotation of the axis during a measurement should be uniform. 
The fulfilment of this condition was easy on account of the delicacy of the galvanometer, 
which required the passage of a current during only a small fraction of a second. Uni- 
formity of rotation during the passage of the current, I obtained by attaching a heavy fly- 
wheel to the apparatus by which the rotation was produced. 
Of the arms of the bridge two, a6, and a d, consisted of coils of thin copper wire 
(silk coated) of exactly equal resistance. The equality of their resistance was tested by 
many careful experiments. The wires were so coiled on the same bobbin as to exclude 
all induction phenomena. The third arm, bc, contained an electrolytic cell, E,, and a 
portion, ¢ e, of the stretched wire, ef, whose length and the length of whose segments, 
ec, cf, and whose resistance per unit length, were known. The cell, E,, consisted of a glass 
tube of nearly 7.5 cm. diameter, with end pieces of plate glass. The side of the tube was 
bored and a tube cemented to it so as to permit the temperature of the liquid in the cell 
to be observed while the cell was sunk in a trough of water. Platinum electrodes were 
firmly cemented to the inner surfaces of the end plates. They were circular, of exactly 
the same size, and of about 7 cm. in diameter. I found it unnecessary to platinize them 
after the manner of Kohlrausch and Grotrian. Platinum wires, welded to the electrodes, 
passed from them through the glass plates, and were covered with india rubber tubing. 
The fourth arm « d, consisted. of an electrolytic cell, E, a box of resistance coils, 
and a portion, c f, of the stretched wire ef. The cell, E,, was of the same construction as 
E,. It was also of the same dimensions, except that its length was considerably less. 
Special care was taken to make all four electrodes of exactly the same area. Contact at c 
was made by pressure. With this arrangement of the wire, e f, however, bad contact at c 
could not affect the result. Contact at a, b, and d was made by means of mercury pools. 
A contact key, #, was inserted in the branch a c, containing the Daniell’s cell. The elec- 
trolytic cells, resistance coils, etc., were joined up in the proper arms of the bridge by 
stout copper wires whose resistance might be neglected. The electrolytic cells, E, and E;, 
were placed side by side in the same trough of water, and the wires were so carried from 
one part to another of the apparatus that no electro-magnetic effect was produced on the 
needle of the galyanometer. The wheel work by which the commutators were driven, is 
not shown in the diagram. With the Daniell’s cell, as source of currents, it was not neces- 
sary that the rate of rotation should be the same for different observations. As a rule, 
however, I used about 250 currents per second. 
I have already described the effect produced on one of Thomson’s galvanometers in 
the galyanometer branch of a Wheatstone’s bridge, when an electrolytic cell is in one of 
the arms, and no precautions are taken to reduce the polarization.* Even if the resistances 
of the arms are of such relative magnitudes, that were they wholly metallic, no current 
would flow through the galvanometer, the polarisation which the passage of the current 
at once sets up in the electrolytic cell introduces a new electro-motive force into the cir- 
cuit which destroys the equality of the potentials of the points bandd; and consequently 

* Proc. R. S, Edin., Session 1874-5, p. 551. 
