206 SCIENCE PROGRESS 



bulk concentration which the solution has experienced. In 

 both methods, the diminution in concentration of the solution 

 was determined by measuring the interfacial tension between 

 the oil and solution, and then reading off the concentration 

 change from the curve of fig. i. This procedure was adopted 

 as surface tension in the case of glycocholate solutions is a very 

 accurate means of estimating the content of the solutions. 

 The same is true of methyl-orange and of congo-red solutions. 

 In spite of the delicacy of the analytical means, however, the 

 actual change in concentration was so small that the value 

 obtained is not accurate to more than 25 per cent. In the 

 case of the inorganic salts examined the tension is no longer 

 sufficiently accurate, and the change in concentration was 

 determined by the usual chemical means. In these cases the 

 result is only accurate to the order of magnitude. 



The emulsion method and the large- drop method of deter- 

 mining the mass adsorbed gave concordant results. A single 

 case may be quoted. 500 ccs. of a 0*318 per cent, solution 

 of sodium glycocholate was vigorously shaken with a known 

 quantity of oil until a homogeneous emulsion was produced. 

 The average radius of the droplets was found to be 4-25 x io~ 6 

 cm. The total adsorbing surface is, therefore, 31,550 cm. 2 . 

 By measuring the interfacial tension of the solution before and 

 after emulsification it was found that the fall in concentration 

 amounted to 0*023 P er cent. Hence the mass adsorbed per 

 cra. 2 = P = 3*6 x io -6 gram. When the bulk concentration 

 prior to emulsification was 0*2 per cent., the quantity adsorbed 

 amounted to 47X io~ 6 gram. The quantity calculated by 

 Gibbs' equation is 5*5 x io -8 gram. There is, therefore, a very 

 great discrepancy between calculated and observed values. 

 It was on this account, indeed, that the second method of 

 measuring P directly was resorted to. With this second 

 method, however, the observed values of p lay between 3 and 

 5 x io -6 gram. There is, therefore, no question of the real 

 existence of the discrepancy. With congo red, methyl-orange, 

 and caustic soda a similar discrepancy exists, the observed 

 adsorption being 20 to 80 times greater than that calculated. 

 In the case of caffeine, however, at an oil surface there appeared 

 to be approximate agreement between the results, though it 

 is to be remembered that the change in concentration is so 

 slight that no great accuracy can be claimed for the observed 



