52 J. G. MACGREGOR ON THE 
different. Assuming them to be different, let P be the polarization of the anode of A and 
n P that of the kathode of B, the current being supposed to pass from A to B. Let mth of 
the current pass through the other electrodes. Then the polarization of the anode of B 
will be F and that of the kathode of A,2¥. Hence the polarization of A will be P—2®, that 
of B,£—n P. They are therefore unequal; and it is easily seen that the effect must be rever- 
sible. (2). It is difficult in cementing Platinum foil to glass plates to prevent entirely the 
penetration of liquid between the foil and the plate. Hence the backs of the electrodes 
which were attached to the partition were probably connected by means of electrolyte. A 
part of the polarizing current therefore passed from a, to b, through the electrolyte. It must 
result that while a,, the face of a,, the face of b,, and b, may be all equally strongly 
polarized, there is a smaller polarization on the backs of a, and b,. If we have a hydrogen 
polarization on the face of a, it is an oxygen polarization which is on the back; and in that 
case it will be an oxygen polarization which is on the face of b,, and a hydrogen polariza- 
tion on the back. Now the experiments of Root * have shewn that hydrogen passes 
very quickly through very thin platinum foil. Hence the hydrogen on the back of b, quickly 
finds its way to the face and affects its polarization while the oxygen on the back of a, not 
having such facility of penetration, does not affect so strongly at least the polarization of 
the face of a. We have therefore a reversible effect. 
Whether the observed effect is due to either or to both of these causes, it was found 
necessary to prevent the passage of the polarizing current from a, to b, by any other con- 
ductor than the wires a,mnb,. 
A second electrolytic cell was therefore made. Like the last it was made of glass 
cemented with marine glue, but the glass partition was cemented in so as to be water- 
tight. The section of the trough was about two square inches. The electrodes were of stiff 
sheet platinum, which were more convenient than those of foil because they could be so 
much more readily cleaned. With this cell I obtained excellent results for a considerable 
time. No reversible effect shewed itself at first. But even before the partition became 
leaky, while the one compartment could still be kept full of water, the other remaining 
apparently dry, it began to manifest itself, indicating that the electrolyte was getting in 
between the cement and the glass. In all cases the appearance of the reversible effect 
was followed by the partitions becoming leaky. 
While the partition had been newly cemented in, the trough was perfectly tight; and 
while it was in that state, I made a sufficient number of experiments to settle the question 
under investigation. Great difficulty was experienced, as usual, in getting the platinum 
plates electrically similar, and many experiments were brought to nought by electrical 
dissimilarity of plates. But now and again it would happen that the plates were obtained 
in practically identical electrical condition, so that the electrolytic cells did not play the part 
of galvanic cells, and in these cases measurements were made. I may quote one or two cases : 
I.—Platinum electrodes in dilute sulphuric acid (density about 1.5). Ratio of distance 
of electrodes in A to distance in B= 6 : 1. Resistance of A and B large compared with the 
rest of the resistance in circuit (viz. the battery). Current had been sent through electro- 
lytic cells from one Grove’s cell for 10 seconds. 

* Pogg. Ann., Bd. CLIX (1876) p. 416. 
