of Gibbs's Theory of Surface- Concentration. 489' 



Discussion of Results. 



The conclusions to which the foregoing experiments point 

 are : — 



(1) Caffeine in aqueous solution, in all probability obeys 



Gibbs's law quantitatively. 



(2) Ordinary inorganic salts — Potassium chloride, Silver 



nitrate, Barium chloride, and Copper chloride — are 

 adsorbed in quantities which are of the same order 

 as the Gibbs's calculated effect, though the experi- 

 mentally found values are in all cases greater than 

 the calculated. Caustic soda shows a more marked 

 discrepancy between calculated and observed values. 



(3) Complex organic salts — Sodium glycocholate, Sodium 



oleate *, Congo red, and Methyl orange — show a 

 very large discrepancy between observed and calcu- 

 lated values. The first thing is therefore to consider 

 the various possibilities of accounting for this large 

 discrepancy. 



The Adsorption of Sodium gli/cocliolate, fyc. 



First, can the excess exhibited by these substances be 

 explained by an extension of the capillary theory ? 

 The expression 



-p g da 



~ RTdc 



is deduced on the assumption that we are dealing with one 

 component only which surfers surface-concentration, i. e., the 

 solute. Since the above-mentioned substances are salts, they 

 will be dissociated, and almost entirely so, at the dilution 

 worked with. Milnerf, in deducing the Gibbs's expression, 

 has allowed for dissociation b}- the introduction of 

 van 't HoiFs factor (i) thus : 



-p _ c da 

 ~T\iTdc 



This is, however, really the expression employed in the 

 foregoing experiments, since in all cases for the evaluation 

 of R (the gas constant) the molecular weight of the solute in 



* Reference will be made later to this substance, 

 f Milner, Phil. Mag., Jan. 1907. 



