318 BULLETIN OF THE UNIVERSITY OF WISCONSIN. 



numbers also vary considerably with the concentration. This 

 at first seems to justify perfectly the position taken by Ar- 

 rhenius. However, the latter has not applied the explanation 

 to all the salts of the magnesia series, as he ought to, but he 

 has simply assumed polymerized molecules in the case of those 

 salts that did not behave according to his theory. He ought 

 to assume polymerized molecules in the case of MgCl 2 , for in- 

 stance, for Hittorf found the migration numbers of this salt 

 strongly dependent upon the concentration, as he did those of 

 MgS0 4 . But it happens that to assume polymerized molecules 

 in the case of MgCL would be very inconvenient for Arrhenius' 

 theory, as his calculations of the factor i show, and so he did not 

 make the assumption which he, in order to be consistent, ought 

 to have made. In the case of CaCl 2 , BaCl 2 , Ca (ISTOg^, 

 Ba(]Sr0 3 ) 2 , Hittorf likewise found the migration numbers 

 strongly dependent upon the concentration, yet Arrhenius does 

 not assume polymerized molecules for these, as he ought to, to 

 be consistent, for these salts agree better with his theory 

 when such an assumption is not made. To assume polymeri- 

 zation of the molecules of such salts as MgCl 2 , CaCl 2 and BaCl 2 

 would lead to the greatest difficulty also in harmonizing the 

 freezing-point results obtained by Jones and Chambers 1 with 

 the dissociation theory ; for they have f ound that the molecular 

 lowering shows a minimum between the strengths 0.1 and 0.2 

 normal, and that in concentrated solutions the lowering is as 

 great or greater than the theoretical lowering, if the com- 

 pounds were completely electrolytically dissociated. Jones and 

 Chambers seek to harmonize their "abnormal" freezing-point 

 lowerings by assuming that the salts form hydrates in the so- 

 lutions. There can hardly be any doubt that the salts in ques- 

 tion unite with water to form hydrates ; but when the authors 

 seek to use this to explain the "abnormally" low freezing-points 

 of solutions, it might be well to remind them that the dissociation 

 theory itself was promulgated to explain "abnormally" low freez- 

 ing-points of solutions, and that since they have found the theory 



1 Amer. Chem. Jour. 23, 89 (1900). Compare also in this connection Chambers 

 and Frazer, Ibid. 23, 512 (1900). 



