of the Theory of Solutions. 477 



very simple phenomenon of affinity, namely the attraction for 

 water. 



Mitscherlich (Lehrbuch der Chemie, 4th edit. 1844, p. 565) 

 had already inquired into the magnitude of the attraction 

 which Glauber's salt has for its water of crystallization, and 

 perceived a measure of it in the diminished pressure of the 

 water of crystallization. He says (loc. cit.), " If Glauber's 

 salt be introduced into a Torricellian vacuum at 9°, the mer- 

 cury sinks 2*5 lines (5*45 millim.) by the liberation of aqueous 

 vapour. On the other hand, water itself causes a fall of 

 4 lines (8*72 millim.). Hence the affinity of sodium sulphate 

 for ts water of crystallization corresponds to the difference, 

 1*5 line (3'27 millim.), i. e. about ^ of a pound (^ kilog.) 

 per square inch (2*615 square centim.)." 



This value, ^h^ of an atmosphere, seemed to me excessively 

 small, and I still retained the opinion, which I had obtained 

 from Helmholtz's Faraday Lecture, that even the weakest 

 chemical forces are exceedingly great. It was therefore 

 desirable to see if this attraction for water could be more 

 directly measured in still simpler cases ; in aqueous solution, 

 for example, we have probably a much simpler phenomenon 

 than the attraction for water of crystallization. Whilst pon- 

 dering over these questions I met my colleague de Vries and 

 his wife as I was leaving the laboratory. De Vries happened 

 to be occupied with osmotic experiments, and called my atten- 

 tion to Pfeffer's determinations. 



You are all familiar with the apparatus which Pfeffer used.. 

 It consisted of a porous pot, the wall of which was rendered 

 semipermeable by precipitating in it a layer of copper ferro- 

 cyanide. This semipermeable membrane, which allowed 

 water but not a dissolved substance, for instance sugar, to 

 pass through, admitted of the measurement of the osmotic 

 pressure, which amounted in the case of a 1 per cent, sugar- 

 solution to two thirds of an atmosphere. 



This pressure was exceptionally great when compared with 

 Mitscherlich'' s number, but there is nevertheless a relationship 

 between them. 



If in fig. 1 we suppose the sugar- Fio-. 1. 



solution contained in the lower left-hand 

 side and separated from the water on 

 the right side by means of a semiper- 

 meable membrane, then the water will 

 pass from right to left until the pressure 

 reaches the amount observed by Pfeffer. 

 In the space above the liquids, however, 



