MEASUREMENT OF OSMOTIC PRESSURE 559 



ment of water through the membrane, would produce effects 

 which might last for a very long time. It was therefore 

 necessary to form the membrane on the inner wall of the cell, 

 an effect which could easily be produced by diminishing the 

 diameter of the pores. When, however, the texture was too fine 

 the membrane was not satisfactory, probably because it was 

 rooted with insufficient firmness to adhere properly to the 

 surface : it was also difficult to develop a good membrane on a 

 cell from which an old membrane had been cleaned off. 



The first step in preparing the cell was to displace the air in 

 the pores by water. This was effected by " electric endosmose." 



Fig. 5. — A r-space above cells. 



The cell was filled with a 0*005 normal solution of potassium 

 sulphate and was then immersed in a similar solution to 

 the lower edge of the glazed portion. By passing a current 

 inwards through the cell, water was drawn in continuously : 

 the cell was then rinsed and soaked and the same process 

 repeated with distilled water. In the later work, lithium 

 sulphate was substituted for potassium sulphate, as it was 

 found that " the quantities of water carried through the porous 

 walls of a cell, under identical conditions, are inversely propor- 

 tional to the relative velocities of the various kathions divided 

 by their respective valencies." 



To deposit the membrane, the cell was set up with a 



36 



