Theory of Electrolytic Dissociation. 219 



95, whereas the constant for water is about 80. It is known 

 that strong acids conduct exceedingly well in aqueous solu- 

 tion, but it has been shown that in liquid hydrocyanic acid 

 they are very poor conductors*, [n general, salts conduct 

 much better in water than in liquid hydrocyanic acid. 

 There are some exceptions to this, however; so, for in- 

 stance, the salts of potassium tested conduct better in liquid 

 hydrocyanic acid than in water. We have here a striking 

 illustration of what is found continually, namely, that whether 

 a solution will conduct electrolytically or not depends upon 

 the specific character of both <olute and solvent, and is quite 

 independent of the dielectric constant of the solvent. 



The approximately additive properties which some solution.- 

 exhibit has frequently been regarded as evidence in favour 

 of the dissociation theory. Additive properties, however, are 

 known to occur where electrolytic dissociation is out of the 

 question, and hence these properties cannot be used as an 

 argument in favour of the Arrhenius hypothesis. In the 

 ease of solids, the molecular heat is approximately equal to 

 the sum of the atomic heats ; in liquids, insulators as well as 

 electrolytes, the molecular volumes and molecular refractions 

 are approximately equal to the sum of the atomic volumes 

 and atomic retractions respectively, and yet who would 

 assume dissociation in these cases on that account? 



It the dissociation theory be held, all the properties of a 

 solution — physical, chemical, and physiological — must be 

 ascribed to the properties of the ions and the undissociated 

 molecules ; and if the dissociation be assumed as nearly com- 

 plete, the properties exhibited by such solution are to be 

 ascribed largely to the properties of the ions. So the colour 

 of aqueous electrolyte- has been ascribed to the colour of the 

 ions. Copper ions are blue: nickel ions are green; cobalt 

 ions are red. &c. But I have shown that benzene solutions 

 of copper oleate. nickel oleate. and cobalt oleate are also 

 bine, green, and red respectively, and that these solutions are 

 a- good insulators as benzene itselff. Furthermore, the 

 cobalt-oleate solution turns blue when heated, and on cooling- 

 turns red again, exactly as aqueous conducting solutions of 

 cobaltous salt<. It lias, therefore, been demonstrated that 

 the colour of solutions is independent of their power to con- 

 duct the current, and cannot be used as an argument in 

 favour of the dissociation hypothesis. 



The reactions in aqueous solutions so commonly used in 



* Coinparo Kahlenberjr and Schlundt. Jour. /Ju/s. Chan. vi. p. 447 

 (1902). 



♦ Jottr.pkys. Chem. vi. y. I (1902). 



