490 Mr. W. C. M. Lewis : Experimental Investigation 



solution has been employed. By simply making this allow- 

 ance for dissociation, therefore, the discrepancy between 

 observed and calculated values for adsorption is still very 

 marked — as a matter of fact, such a correction acts in the 

 opposite direction. 



The next possible explanation may lie in incorrectness in 

 the assumption (which has been made throughout) that the 

 water suffers no surface density-change. 



It is very probable that water has not quite the same 

 density in the surface-film as in the bulk of the liquid. 

 Evidence for this is to be found in the phenomenon of 

 heat evolution observed by Parks and others, when water 

 is poured upon finely- divided powders of silica, quartz, 

 glass, &c, where chemical action is of course excluded. 

 Also one must not overlook the fact that in the adsorption of 

 sodium glycocholate and the dyes we are dealing with surface- 

 concentrations far exceeding the ordinarily accepted values 

 for their solubility. Of course in using the term " solubility" 

 we assume that the surface-layer water possesses the same 

 properties as water in bulk. For a solution of sodium 

 glycocholate of 0*25 per cent, bulk concentration, V was 

 found to be 5 x 10 ~ 6 grm./cm. 2 The value of the range 

 through which the surface-Concentration is sensibly different 

 from the bulk has been shown to be approximately 14 x 10~ 6 cm. 

 Hence the average surface-concentration is 0*37 gram per 

 c.c. or 37 per cent., while the ordinarily accepted value for 

 the solubility of sodium glycocholate is 3*9 per cent. There 

 was, however, no visible colour change on the oil surface 

 due to any precipitation. 



Again, methyl orange * showed an adsorption for which 

 F = 5*5 x 10~ 6 , whence the average concentration is 39 per 

 cent. The actual solubility is only 0'078 per cent. 



The substances therefore which have shown very great 

 discrepancies as regards Gibbs's theory are those ichose surface 

 concentration greatly exceeds their solubility in the solvent- 

 For the other substances examined, viz. caffeine and the 

 inorganic salts, in no case is the ordinary solubility exceeded, 

 and the adsorption of these substances is in much closer 

 agreement with theory. Thus the surface-concentration of 

 caffeine was found to be 0*26 per cent., while its solubility 

 at ordinary temperature is 1*35 per cent. For silver nitrate 

 the surface-concentration was approximately 0*28 per cent., 

 while its solubility is 122 per cent, at 0° 0. Similarly for 

 potassium chloride, barium chloride, and copper chloride. 



* See previous paper, I. c. 



