342 
PROFESSOR 0. MASSON ON IONIC VELOCITIES, 
suitable length and considerable capacity is marked with a file at the place where 
these (or other) indicators should meet. It is then filled with the KCl jelly and 
connected with the cells, in hath of which is placed coj^per sulphate solution. A 
copper anode and a platinum cathode are used. During the progress of the experi¬ 
ment, two boundaries travel as usual, though only one is visible, viz., that between 
KCl and CuCL. It will, however, meet the other, viz., that between KCl and K2SO4, 
at the mark previously made ; and the circuit must be broken when the blue boundary 
reaches this point. The cathode will have gained in weight by as many miUigramme 
ecpiivalents of copper as there were milligramme molecules of KCl originally present 
in the whole tube ; and this may l)e taken as a test of the correctness of the 
experiment. That it is so, is evident from the fact that there will have travelled 
across that section of the tube, where the meeting of the margins occurred, all the 
K ions originally present on one side of it, and all the Cl ions originally present on 
the other, but nothing else. These are together equal in number to the total K, or to 
the total original KCl molecides. The tube is cut in two at the mark as soon as the 
experiment is over, and the contents of each part are then analysed separately. Cl and 
Cu being estimated in tlie part nearest the anode, and SO4 and K in the other. The 
results may he checked by estimating the Cl that has escaped into tlie anode cell, and 
the K that has escaped into tlie cathode cell. The residts give the value of n directly 
for each end of the tube, with which the value of the original n may be compared. 
The relative specific velocities {ujv) for Cu and Cl, and for K and SO4 follow from con¬ 
sideration of the exchanges at the anode end and the cathode end respectively; and, 
as those of K and Cl are already known, as determining the meeting point, the 
s})ecific velocities of all four ions may be compared with one another. 
One such experiment has been carried out, but the accuracy of the analytical 
results was, to some extent, spoilt by unforeseen difficulties that cropped up in the 
course of the work, due in part to the presence of gelatine in the solutions of the 
tube-contents, and in part to the formation of cupro?fS chloride in the anode cell and 
on the anode. This formation of cuprous chloride has since been observed in other 
experiments. It may be seen on tlie anode as a white crust after it has been washed 
with water, alcohol, and ether ; and its formation there causes the anode to lose less 
than the calculated weight instead of more, as is usual. It is intended to repeat the 
experiment described aljove, taking all precautions to ensure accuracy. In the mean¬ 
time it may be said that the residts, though rough, tended entirely towards the 
confirmation of the theory. Thus they showed, as might be expected, that very little 
change of concentration or of K velocity occurs across the KCI/K.2SO4 boundary, but 
tliat at tlie blue boundary the Cl approximately doubles its velocity and reduces 
its concentration to about two-thirds. These figures (which are only rough 
approximations) agree, according to the formulae already given, with a chlorine 
transport number [p') in CuCb of ‘ 67 , taking its value (y>) in KCl as ' 5 . The author 
is net aware of any Hittorfian experiments with CuCL ; but, to judge from those 
