9 



of colloids. The well-knowTi osmometer method m which the 

 membrane is not always strictly semipermeable involves a further 

 very serious and not completely evaluated complication described 

 in Donnan's approximate theory of membrane equilibria. Results 

 of such measurements therefore are almost always subject to a certain 

 ambiguity of interpretations. The classical methods of boiling 

 point, lowering of vapour pressure, and freezing point lead to results 

 which are beyond question, only provided that the essential conditions 

 for the appUcation of these methods are fulfilled. 



The classical experiments of Krafft and of Smits using boiling 

 point and vapour pressure respectively were completely vitiated by 

 the unsuspectedly large amounts of dissolved air from which soap 

 solutions caimot readily be freed and which develop a partial pressure 

 quite comparable with the lowering of the vapour pressure of an 

 ordinary crystalloid or electrolyte. To carry out a single vapour 

 pressure measurement reaUy requires several weeks of effort. The 

 result then obtained is in agreement with those by the dew point 

 method, to be described below. The osnometer data of Moore and 

 Parker, 1^ in tlie case of soaps were wholly erroneous on account of the 

 effects of carbon dioxide. 



The effect of dissolved air can be completely eliminated and 

 accurate measurements of the vapour pressiire made, by a modification 

 of the dew-point method described by McBain and Salmon. 



The freezing point method is applicable where solutions remain 

 in sol condition at the freezing point and where ice separates out in 

 crystals free from colloid. In the case of soap solutions the latter 

 condition was attainable by inoculation, but very few soap solutions 

 are liquid at 0° C. Determinations of all these have been published 

 by McBain, Laing, and Titley.^* 



Thus the bulk of the available data for soap solutions are those 

 obtained by the dew-point method which possesses an additional 

 advantage in that it can be utilised at any temperature. 



The results of the lowering of dew-point and freezing point agree 

 in establishing the fact that soap solutions have a very real osmotic 

 pressure of the same order of magnitude as that of a non-electrolyte 

 such as sucrose. A continuous series of values for the osmotic 

 activity is obtained, depending upon the concentration and position 

 in the homologous series, which range from a fraction of that for 

 non-dissociated crystalloid to that of a moderately dissociated 

 electrolyte. 



In discussing osmotic data, it is almost always essential to keep 

 clearly in view the very important effect which hydration exerts in 

 magnifying the apparent osmotic pressure more especially in con- 

 centrated solution. For instance, the apparent dissociation of 

 electrolytes frequently exceeds 100 per cent, according to osmotic 

 methods. 



It is generally agreed that hydration is greatest at lower tempera- 

 tures. In the case of the potassium salts of the lower fatty acids this 

 enhanced osmotic effect is clearly visible at 0° as compared vidth the 

 more normal results at 90°, but in spite of the magnification of osmotic 

 data at lower temperatures the osmotic activity exhibited by soap 



f 



