206 PROCEEDINGS OF SECTION B. 
electrode (a copper one) which is to receive the current, in the 
conventional sense, from the battery, and which will therefore 
act as the anode in the electrolysis, is filled up with a cupric 
sulphate solution; while the other flask, which contains a 
platinum cathode, is filled with a solution of potassium chromate. 
At the moment the circuit is closed the tube contains no ions 
except those of the original salt, say, potassium chloride, and 
the composition of the jelly and the potential slope are uniform 
from one end of the tube to the other. The result is the start- 
ing of a procession of potassium ions towards the cathode, and 
of an opposite procession of chlorine ions towards the anode, 
each kind moving with a velocity determined by its own specific 
velocity coefficient, by the degree of ionisation of the salt, 
which varies with the concentration, and by the potential slope. 
Since ionisation and potential slope are both necessarily the 
same for both K and Cl and for all parts of the K Cl jelly, it 
is clear that an observation, if practicable, of the actual speeds 
with which the two processions move simultaneously in opposite 
direction will give a fair measure of the relative speeds of the 
two ions under any equal conditions, 7.e., of their relative specific 
velocities. 
Now it is obviously necessary that, as the rear end of the K 
procession vacates the part of the tube nearest the anode, it 
shall be followed up by the only other available cations, namely, 
Cu ions from the copper sulphate solution; else the current 
would be arrested. In the same way, the rearmost Cl ions, at 
the other end of the tube, must be followed up by Cr O, anions 
from the potassium chromate solution. The tube thus im- 
medi..tely becomes differentiated into three parts—a colourless 
K Cl jelly in the middle, a blue Cu Cl, jelly at the anode end, 
and a yellow K, Cr O, jelly at the cathode end; and the first 
of these continually decreases in length as the others grow 
towards an ultimate meeting point. The boundaries between 
the coloured and colourless portions, while advancing, remain 
sharply defined. In other words, there is no mixing up of the 
ions where Cu follows K, nor where Cr O follows Cl; and it 
can be shown theoretically that there should be none, provided 
that (as in this instance) certain conditions are fulfilled. The 
rates of advance of these boundaries can be accurately measured 
by means of a stop-watch and the graduated scale of the tube ; 
and they afford the required test of the actual speeds of the K 
and Cl ions respectively, which are thus rendered visible by 
contrast with their coloured followers. The relative speeds, and 
the consequent meeting point of the coloured boundaries, are 
calculable with fair accuracy from the first few readings; but 
the experiment is continued till this meeting actually occurs. 
Of course, the rate of advance of the blue boundary is a 
correct indication of the actual velocity of the Cu ions as well 
