234 Wisconsin Academy of Sciences, Arts and Letters. 
ment performed at room temperatures showed a very rapid rise 
of liquid in the osmometer tube,—see pressures measured in 
the quantitative measurements described below. 
28) The inner liquid was a saturated solution of Li Cl in 
pyridine, the outer liquid pyridine and the septum rubber. 
The liquid rose in the osmometer and Li Cl did not pass 
through the septum into the outer liquid in appreciable quan¬ 
tity. 
Experiments 23 to 27 show that vulcanized caoutchouc is a 
‘ ‘ semipermeable ’ ’ membrane when it separates sugar solution 
in pyridine from pure pyridine. The data lead one further to 
the conclusion that the gas laws do not govern the phenomena, 
which the measurements to be detailed later confirm. No. 28 
shows that when Li Cl is used as solute in an otherwise similar 
experiment, the rubber again acts as a semipermeable septum. 
The reasons for this behavior have already been discussed in 
connection with the AgN0 3 solutions. 
29) The inside liquid was 0.1 normal AgN0 3 in pyridine, 
the outside liquid a saturated solution of cane sugar in pyri¬ 
dine, the membrane being rubber. The level in the osmometer 
fell, showing the current to be from the AgN0 3 solution to that 
of the sugar. On further examination it was found that 
AgN0 3 had also passed into the sugar solution but in small 
amount. 
30) This experiment was identical with No. 29 except that 
parchment was used as the septum. The test was continued 
for 3 days. No change was observed. 
In the light of what has already been said, it is clear that 
the results in Nos. 29 and 30 are such as might have been an¬ 
ticipated. 
31) The inner liquid was a solution of FeCl 3 in toluene, the 
outer toluene, the membrane rubber. The main direction of 
flow is toward the solution, but FeCl 3 also passes through the 
membrane in considerable quantity, which was to have been 
expected since this salt is soluble in hydrocarbons. The FeCl s 
gradually disintegrates the septum. 
32) This experiment was identical with that of No. 31 ex¬ 
cept that iodine was used as the solute instead of FeCl 3 . 
