114 



A. P. MATHEWS 



power of many salts for the globulin of the hemp seed edestin. I 

 append their results here to show how far they agree with the theoreti- 

 cal deductions which were worked out on p. 102. In Table 9 E c E a 

 represents the difference between the ionic potentials of the salts; 



TABLE 9. 



the third column represents the number of c.c. of a normal solution 

 of the salts which will just dissolve one gram of edestin. I have 

 taken these figures from the chart given by Osborne and Harris. They 

 are only approximate. The theory requires that the more negative the 

 difference E c E a , the smaller the amount of salt necessary to dis- 

 solve a given amount of the edestin. It will be seen that this relation- 

 ship holds very well if we compare the iodides, chlorides, and bromides 

 of sodium, lithium, and potassium respectively. Unfortunately, 

 the great uncertainty of the solution tensions and ionic potentials 

 of sodium, potassium, and lithium, prevent quantitative comparisons 

 between the different metals. In the fourth column under K, I 

 have computed the constant by the formula on p. 102, using instead 

 of the logarithm of the dissolution the logarithm of the ratios of 

 number of c.c. of the different solutions necessary to dissolve one 

 gram. The figures are placed between the salts compared. The 

 formula used was: 



i , c.c. T 



c.c., 





Osborne and Harris draw the conclusion that the solubility 

 is independent of the nature of the base, but I think their figures 

 speak for themselves. The differences between potassium, lithium, 

 and sodium are small, to be sure, but nevertheless apparent. The 

 importance of the base becomes obvious so soon as any other salts 

 are examined in which the base has a higher ionic potential than 



