332 rEPORT—1890. 
formed, through which, according to the observations of Traube, no copper 
salt can diffuse, we are led to a somewhat strange question. The fact that 
no copper salt can pass through the membrane is evidence that the copper 
ions existing in the salt solution are likewise unable to pass. But as the 
electricity in electrolytes travels only with the ponderable ions, we are 
met by the alternative either that the refusal of the copper (and ferro- 
cyanide) ions to pass through the membrane will cause the current wholly 
to stop, or that the electricity will deposit the copper ions on the mem- 
brane and itself alone pass through. The semipermeable membrane must 
in the first case act as an insulator; in the second case it must act as 
a metallic diaphragm. Both these cases are so unexpected that the 
described experiment at once acquires a special interest. 
By performing the experiment we find that the second alternative 
holds good. The current becomes rapidly weaker, and after ten minutes 
we can easily observe a very marked polarisation current in inverse direc- 
tion to the primary current. After some hours of current the parchment 
paper containing the semipermeable membrane on the positive side is 
coated with a layer of metallic copper, and this is evidence that the copper 
zons are filtered off by the semipermeable nembrane. 
From this experiment it follows that the semipermeable membrane 
really acts as a sieve, not only as regards compounds, but also for ions, 
allowing some of them to pass and retaining others; for we know, for 
example, that potassium chloride can pass the membrane of copper prus- 
siate, and therefore the ions K and Cl do so, while barium chloride and 
potassium ferrocyanide are retained. In the two last-mentioned cases one 
of the ions has the power of passing, but is retained by the other. At the 
first moment the Cl ions of the barium chloride will of course go through 
the membrane, while the barium ions stay behind. But by this separation 
@ separation of positive and negative electricity also takes place, and 
thereby forces will arise tending to draw the Cl ions back. Finally a 
double layer of electricity is formed, causing a potential difference on both 
sides of the membrane, whose value depends only upon the molecular 
concentration of the electrolyte, and in no way upon its nature. 
If the formation of the double layer is prevented, free diffusion of the 
passing ion takes place. By adding to the barium chloride some salt 
whose metal can pass through the membrane—for instance, some salt of 
potassium—the Cl ions at once will traverse the membrane, but the same 
number cof K ions must go along with them. In this case, however, it 
may be assumed that the added potassium salt undergoes a double decom- 
position with the barium chloride, forming potassium chloride, which is 
able to diffuse through the membrane. But we can also cause the Cl ions 
to pass by putting some diffusible negative ions on the outside of the — 
membrane—for instance, copper nitrate. Then we soon find chlorine — 
outside and a nitrate inside the membrane. In this case it is impossible — 
to assume a double decomposition, because both the salts are separated by 
the membrane, which prevents the diffusion of the barium chloride, as 
well as of the copper nitrate; and the explanation, by taking into account 
free migrating ions, seems to be the only sufficient one. 
$ 
The above-mentioned double layers and potential differences, occurring — 1 
at semipermeable membranes, when one of the ions of the electrolyte is 
retained, are probably the source of the potential differences and currents 
we meet with in living matter, because the cells of organisms are all coated 
with such semipermeable membranes. It is perhaps not too rash to hope 
