258 Wisconsin Academy of Sciences, Arts and Letters. 
outer liquid—but rather with a solution more or less dilute. 
When the apparatus is at rest, the slow diffusion processes, 
aided by gravity perchance, tend to keep the outer liquid of 
uniform concentration; but these agencies clearly can not be 
relied upon to keep pure solvent in contact with the outer sur¬ 
face of the membrane in osmotic pressure measurements. In¬ 
deed, when the surface of the membrane is not smooth and 
presents considerable area to the outer liquid, a film of liquid 
is apt to adhere tenaciously to the outer surface of the septum 
in spite of the effects of diffusion. In the usual osmotic ex¬ 
periments using aqueous sugar solutions in a cell made by pre¬ 
cipitating copper ferrocyanide on the inner side of an unglazed 
cup, the slight amount of sugar that passes through very likely 
lingers very tenaciously in the pores of the cup just outside of 
the actual membrane, forming there a film of solution of such 
strength that its effect upon the osmotic pressure is not a neg¬ 
ligible quantity. At least its effect can not be assumed to be 
negligible without further experimental work. Nor would 
stirring the outer liquid in such a case as this be apt to remove 
the difficulty. In the osmotic cells described above where thin 
rubber membranes supported by cloth and perforated steel discs 
were used, the effect in question is no doubt less than in the 
walls of a porous cup, but it is by no means negligible. Here 
the cloth and the disc hinder diffusion, and it is very necessary 
to stir the outer liquid thoroughly and continuously in making 
the osmotic pressure measurements. 
1 would like to emphasize here once more then that it is very 
essential to stir the contents of the osmotic cell and also the 
outer liquid continuously in any attempt to measure directly 
the maximum osmotic pressure that may he produced in a given 
case; and that since in all past experiments this has been en¬ 
tirely neglected , the results of such experiments can not he con¬ 
sidered as final and conclusive. In reality, as has been pointed 
out above in connection with the ether, water, chloroform ex¬ 
periment of L’Hermite, the membrane itself ought also to be¬ 
stirred during the process. This is of course less necessary 
when the septum is quite thin than when it is thick. It should 
be borne in mind, however, that even stirring the contents of 
