EXPERIMENTAL WORK ON OSMOTIC PRESSURE 627 



sequent flow of liquid from end to end secures thorough mixing, 

 and periodically wets the branches through which the current 

 of air passes. Since the air must be passed slowly through the 

 train of vessels, the experimental work is rendered very tedious, 

 and the time consumed in determining the vapour pressure of one 

 solution runs into days. The results are accurate to within 

 5 per cent., and from the lowering of the vapour pressure which 

 is thus ascertained the osmotic pressure P of the solution may 



be calculated by the formula p = — . log;,=r» where p and p l are 



a pi 



the vapour pressures of the water and the solution respectively, 

 5 is the density of water at the temperature of the experi- 

 ment, and a is the vapour density of water under the standard 

 atmosphere A. 



It is very interesting to compare the values of the osmotic 

 pressure of concentrated cane sugar solutions reached in this 

 indirect way with those obtained by direct measurement, and in 

 Table III. the data necessary for the comparison are recorded; 

 the figures are those obtained at o° C. by the Earl of Berkeley 

 and Mr. Hartley : 



Table III 



Osmotic pressure in atmospheres. 

 Gram molecules of cane sugar Directly Calculated from 



per litre of solution. observed. the vapour pressure. 



1-58 67-51 69-4 



1-93 10078 IOI'9 



2-19 13374 I36"0 



The amount of material so far available for a comparison of 

 the two sets of values is not very large ; but the agreement is 

 sufficiently striking, and shows that the direct method of 

 determining osmotic pressure yields values which may be 

 regarded as very nearly correct. It is noteworthy that the 

 osmotic pressure calculated from the vapour pressure is, in all 

 the three cases just cited, higher than the osmotic pressure 

 observed directly. This is only what might be expected, for 

 all the experimental errors associated with the direct deter- 

 mination of osmotic pressure would tend to give too low a 

 value. In so far also as the sugar solutions are compressible, 

 the observed value of the osmotic pressure will be less than 

 the real value. 



The direct determination of the osmotic pressure of a 

 solution is conditional on the discovery of a satisfactory semi- 



