Dr. F. Tinker on Osmotic Pressure. 439 



that pi will be greater than p^. This explains why osmotic 

 flow usually takes place from pure solvent to solution. The 

 pure solvent generates a greater solvent pressure inside 

 the membrane than the solution does. 



We must not, however, suppose that the solvent of necessity 

 always flows from pure solvent to solution. It would flow 

 the other way if for any reason the partial pressure * 

 generated in either the membrane or the vapour phase 

 proper by the solution were greater than the pressure 

 generated by the pure solvent itself. It is not unlikely that 

 solutions of ordinary phenol, which apparently under certain 

 conditions have a partial water vapour pressure greater than 

 the vapour pressure of pure water itself f, come under this 

 category. Prof. Adrian Brown and the present author have 

 proved that the phenolic solutions in general are very 

 abnormal in their osmotic behaviour {. 



In this connexion it is also interesting to note the effect 

 of compressing the solution whilst leaving the pure solvent 

 under atmospheric conditions. It is evident that the 

 shrinkage caused by compression will tend to make e 

 negative, so that ultimately, by the application of a pressure 

 great enough (V 2 — l> 2 -\- e) will become negative. When this 

 occurs pi will become loss than p/ and the solvent will flow 

 from the solution to the pure solvent. This is actually what 

 happens when a pressure greater than its osmotic pressure 

 is applied to the solution. 



Let us now consider the causes for the abnormalities in 

 the phenomena of osmotic flow ; and let us confine attention 

 in the first place to the case in which two dilute solutions of 

 different solutes, having equal concentrations and under the 

 same external hydrostatic pressure, are on opposite sides of 

 the membrane. If p/ and pi' are the solvent pressures 

 generated in the membrane in each case we have from the 

 equation [12] 



Pi~Pi" _ N / Q"-Q' \ ri , n 



where Q' and Q" are the heats of dilution of the two 

 solutions. Similarly, for the vapour pressures proper, we 

 have 



Pi'-Th" 



Pi 



N i'Q"-Q'\ r, .-, 



* I. e. "Partial pressure of the solvent." 

 t Schreinemakers, loc. cit. p. 14- f footnote). 

 X Proc. Roy.'Soc. B. lxxxix. p. 119 (1915), 



