222 Prof. L. Kahlenberg on the 



used by advocates of the dissociation theory were far from 

 dilute ; and, at best, the agreement between calculated and 

 observed values leaves much to be desired. 



Many colloidal solutions coagulate when treated with 

 solutions that conduct electro lytically. Attempts have been 

 made to ascribe this coagulative power to the free ions which 

 the electrolyte is supposed to contain. It is true, however, 

 that the power to coagulate colloidal solutions is by no means 

 confined to electrolytic solutions ; and hence any one claiming, 

 for example, that sodium-chloride solution coagulates colloids 

 because it is an electrolyte, must assume the burden of proving 

 his proposition. 



The theory of electrolytic dissociation is unable to furnish 

 a satisfactory explanation of the phenomena in aqueous solu- 

 tions, but it is in the realm of non-aqueous solutions that the 

 hypothesis proves especially impotent. When the theory was 

 promulgated but few non-aqueous solutions had been examined 

 as to their power to conduct the current, and those few solu- 

 tions, in which ether, alcohol, chloroform, and hydrocarbons, 

 for instance, were the solvents employed, were found to be 

 poor conductors. So the idea that only aqueous solutions 

 conduct electricity to any considerable extent gained ground 

 and held sway for a number of years with remarkable tenacity, 

 the view being strengthened by the fact that in the solvents 

 just mentioned molecular weights as determined by boiling- 

 point, or, where possible, by freezing-point methods, were 

 not " abnormally " low, as in the case of solutions of many 

 ordinary salts in water. But now we know of non-aqueous 

 solutions which are as good and even better conductors of 

 electricity than aqueous solutions ; and these solutions, when 

 tested by the methods of determining molecular weights just 

 mentioned, yield results that are generally higher then those 

 computed from the formulae of the solutes ; only exceptionally 

 is the reverse the case. In the course of my work on non- 

 aqueous solutions, I have pointed out numerous examples of 

 this ; and recently the splendid researches of Walden have 

 established many more. His work on solutions in liquid 

 sulphur dioxide especially shows that in this solvent " abnor- 

 mally " high molecular weights are generally found in the 

 case of solutions of excellent conducting power. The field 

 of non -aqueous solutions is a very large one, and it is here 

 that the theory of electrolytic dissociation can really best be 

 tested. It is here that the fact that the theory is untenable 

 becomes particularly apparent. Whetham's statement* that 

 work with non-aqueous solutions is rendered difficult by the 



* L. c. p. 19. 



