114 
MR. B. D. STEELE ON THE MEASUREMENT OF IONIC VELOCITIES 
sionally be produced even when conditions 1 and 3 are not fulfilled, but, as is seen 
from the e(|uations of Weber, already given, its velocity is no longer that of the 
ion in the intermediate solution. Ciondition No. 4 is sutliciently obvious, neverthe¬ 
less it is necessary to point out that it is not sufficient that the difference in 
density of equimolecular solutions should be known; for, since concentration 
changes are brought about during electrolysis in the sense that the indicator 
solution is always (provided condition 1 is fulfilled) of less concentration than the 
measured solution, it may, and sometimes does, happen that an indicator solution 
when jilaced l)eneath the solution to be measured, and which in equimolecular 
solutions is more dense than the lattei’, becomes, through these changes, lighter 
than and accordingly rises through the overlying solution, with the result that 
no Ijoundary can he obtained ; this behaviour is shown with 2 N potassium chloride 
solution followed by 2 N copper chloride. According to the I’elation cjc = p^p', the 
concentration of the copper solution becomes reduced to 1'2 N, the density of which 
= 1'0711, whilst that of 2‘0 N potassium chloride = 1'0886. Hence for this reason 
copper chloride, although its density, as calculated from Valsox’s moduli (see 
Nernst’s ‘Theoretical Chemistry,’ p. 333), is much greater than that of potassium 
chloride, cannot be used as an indicator for the latter from underneath. 
In addition to the experimental evidence of these concentration changes that has 
been given by Massox, further qualitative evidence is found in the fact that in an 
experiment in which lithium or magnesium follows potassium in a solution of 
potassium chromate, so that there is formed elect roly tically a solution of lithium 
or magnesium chromate overlying one of potassium chromate, the indicatoi- solution 
is seen to be distinctly lighter in colour than the measured. 
Since the colour depends only on the anion, this shows the 
concentration of the latter is less in the indicator solution. 
More direct and conclusive evidence was obtained in the 
Ibllowing manner :— 
An apparatus, shown in tig. 2, was employed. The tubes 
A and B, instead of having the anode and cathode cells 
ground in, are provided with pieces of thick, strong, india- 
ladDber tubing, which are tirmly bound on with copper wire, 
into which the cells C and C' can be fitted. 
It is found in practice that no displacement of the latter 
takes place during the course of an exjjerinient, and, in order 
to reduce the danger of melting the jelly by the current, the 
E necks of the vessels C and C' are made a little larger than the 
tubes A and B; the tube E serves the double purpose of a 
support or handle, by means of Avhich the apparatus may be 
held, and also to allow for any slight contraction or expansion 
that may take place through unequal heating. 
