president's address — SECTION B. 55 



done in the effort to correlate the formation of a conducting solu- 

 tion with certain of the physical properties of the solvent. Of these 

 properties there is one, and only one, the magnitude of which appears 

 to vary with the 'ionising' power of the solvent. This is the 

 dielectric constant, and its possible relation to the power of a 

 solvent to take part in the formation of a conducting solution was 

 pointed out by Nernst and by Thompson, and has been thoroughly 

 investigated by Walden. 



Consideration of the foregoing facts has led to the suggestion 

 that the necessary antecedent to the formation of a conducting 

 solution is the formation of a compound between solvent and 

 solute. I think that there is not yet sufficient evidence to settle 

 the question here set forth, but the following considerations indicate 

 clearly that such compound formation cannot be the only necessary 

 antecedent. 



(1) Hydrogen bromide alone and water alone do not form 

 conducting solutions in liquefied sulphur dioxide, but mixed 

 solutions of these two substances in this solvent do conduct fairly 

 well. This can be explained by compound formation between the 

 water and the hydrogen bromide with formation of an oxonium 

 compound, but cannot be explained by compound formation of 

 either solute with the solvent. 



(2) It might be argued on the last quoted results that the 

 conductivity of a solution of hydrogen bromide in water was to be 

 associated with the prior formation of an oxonium compound, and 

 this is probably correct ; but if this were the only necessary con- 

 dition we should expect to get a much more highly conducting 

 solution of hydrogen bromide in acetone, since there is evidence 

 that a much more stable compound should result in the case of the 

 last named pair of substances. This expectation is not realised. 



(3) Hydrated salts are not all soluble in water, as for example, 

 Ca SO4 2H2O, which is very sparingly soluble ; many other salts 

 which at the temperature of solution do not yield hydrates are 

 freely soluble and form highly conducting solutions. 



A similar lack of parallel between compound formation and 

 solution formation is instanced by the extreme solubility of silver 

 iodide and chloride in liquefied ammonia, and the extreme insolu- 

 bility of calcium chloride, although all three substances readily 

 form addition compounds with ammonia. Moreover, although 

 large numbers of metalhc salts are known with each of the solvents 

 — water, ammonia, sulphur dioxide, and the hydrides of chlorine, 

 bromine and iodine — -it is probable that the majority of them do 

 not dissolve in the appropriate solvent ; and it is therefore certain 

 that the formation of a compound between a sohd substance and 

 a hquid which can function as a solvent does not alone justify the 

 anticipation of a solution being formed, much less a conducting 

 solution. 



The detailed study of a few solvents has brought out many 

 striking analogies in the mechanism of electrolysis in these solvents. 



