II. OSMOTIC PRESSURE MEASUREMENTS 61 



in nature from the solutions bathing the cell in its natural habitat, 

 and equilibrium measurements may be reliable in such cases. Some 

 cells are so perfect as osmometers, their membranes being readily 

 permeable to water but impermeable to other small molecular species, 

 that they have been employed successfully in the determination of 

 the molecular weights of solutes. De Vries was able to determine the 

 molecular weight of rafRnose by plasmolysis experiments with the 

 cells of Tradcscantia discolor. Overton compared the concentrations 

 of solutions of sugars and amino acids that would just cause plasmol- 

 ysis of the cells of Spirogyra and found them to be closely propor- 

 tional to the molecular weights of these solutes. The subject has 

 been reviewed by Lucke and McCutcheon (19) and Luck^ (20). 



Osmometers. The direct determination of the osmotic pressure 

 of a solution for the purpose of estimating molecular weights of 

 solutes, must rely, for dependabihty, upon the use of a membrane 

 which is impermeable to the solute the molecular weight of which 

 is desired, but which is readily permeable to the solvent and to any 

 other solute con.stituents that may be present in the solution. 

 Membranes should be of uniform properties as to thickness and 

 porosity and subject to easy and uniform reproducibility. The ex- 

 tensive investigations of Manegold (22) have furnished much infor- 

 mation on the production of such membranes. Collodion membranes 

 can be readily prepared in the forms required for use in the various 

 osmometers described, and such membranes have been found gener- 

 ally satisfactory for measurements of the osmotic properties of high 

 molecular substances in aqueous solutions. Cellulose membranes 

 regenerated from cellulose nitrate or viscose are obtainable commer- 

 cially (as unwrinkled material that has never been dried or treated for 

 waterproofing) or may be regenerated from cellulose nitrate by a 

 method such as that described by Montonna and Jilk (23). Such 

 membranes have been used successfully for solutions in organic sol- 

 vents. A primary limitation to the use of any such membranes is 

 that the molecular size of the solute under stud,y must be great enough 

 so that the membrane will not allow diffusion of the solute through its 

 pores. If the porosity of the membrane is not known, analysis for 

 the solute in the external phase, at the end of a determination, should 

 always be made to be certain that the membrane used has acted 

 dependably in this regard. Carter and Record (40) describe a 

 method by which commercial nonwaterproofed viscose can be pre- 

 pared as membranes of various degrees of porosity for use in aqueous 



