19 



valerianate is not markedly altered upon the addition of strong acid 

 which liberates the soluble Uquid fatty acid, although excess of 

 base raises the viscosity. In the higher members of the homologous 

 series, where insoluble acids and acid soaps are in question, addition 

 of strong acids increases the viscosity appreciably through formation 

 of coUoidal acid soap ; alkaU does so also, although through formation 

 of ionic micelle and neutral colloid. In accordance with this the 

 higher soap solutions [require a slight excess of alkali to clarify them 

 completely by driving back hydrolysis. The authors named thought 

 to identify the point of complete clarification with the point of 

 minimum viscosity without carrying out the quantitative work and 

 without a knowledge of the constitution of these solutions. 



Lifschitz and Brandt have recently studied refractive index at 

 70°. They find that the refractive index varies Unearly with the 

 concentration; this held good also in the alcoholic true solutions. 

 This undoubtedly shows that the molecular refraction is independent 

 of whether the soap is present as crystalloid or coUoid, and in the 

 latter case is independent also of the actual constitution and of the 

 degree of dispersion of such colloid. This conclusion was du-ectly 

 confirmed by the agreement of the values observed with those 

 calculated from the atomic refractivities which, of course, have been 

 obtamed in the study of crystalloids. 



(f) Ultrafiltration and Dialysis. 



The behaviour to be expected in the ultra-filtration or dialysis of 

 soap solutions is now clear on the basis of the theory which we have 

 established. In the case of ultrafiltration the effect will depend 

 largely upon which fatty acid is taken and the concentration of the 

 soap. In dilute solution or solutions of the salts of the lower fatty 

 acids, thej/ wfil tend to pass through the filter unchanged. In more 

 concentrated solutions of the higher soaps which contain very little 

 crystalloid other than the alkali ions accompanying the ionic micelle, 

 the membrane wiU retain practically all the soap. Intermediate cases 

 will be particularly interesting as they will afford a direct quantitative 

 measurement of the crystalloid matter present in such solutions. 



The only published investigation is the brief note of Mayer, 

 Schaeffer, and Terrione, already referred to, whose results were 

 apparently m quantitative agreement with that here outhned. 

 A systematic investigation is now in progress whose results likewise 

 agree with prediction. Further it is readily sho^\^l that this may be 

 made a general method for the study of hydration of colloids. It is 

 only necessary to include in the solution to be filtered some readily 

 analysed material which is not sorbed by th'e coUoid and then to 

 determine the increase in concentration of this reference substance 

 which the filtrate exhibits. The amount of this increase affords a 

 quantitative measurement of the total amount of water which was 

 abstracted by the' colloid from solution whether as adsorbed, or 

 chemically combined water. 



As regards dialysis, there are two points to be emphasized, the 

 one theoretical, the other practical. In the first place the crystalloid 

 present will be able to i^ass through the membrane so that in no case 



