ON ELECTBOLTSIS. 349 



dilute solutions, where the friction on the water is almost the only thing to be con- 

 sidered. 



But with regard to very dilute solutions, I believe we encounter some great 

 difficulties, at least for certain compounds. 



Ist. There may be double decomposition of the dissolved electrolyte with the 

 impurities of the solvent. In very diluted aqueous solutions these impurities, even 

 in the cleanest water, conduct better than the dissolved solid. It may be that this 

 double decomposition may have no great influence on the results. 



2nd. A greater influence can be exerted by dissociation of the salts. It is 

 known that solutions of sulphate of copper, &c., are acid, that a solution of chloride 

 of magnesia when boiled emits vapours of hydrochloric acid, that chloride and 

 other salts of ammonia are dissociated in their aqueous solutions, &c. The increase 

 of the dissolving water must increase these dissociations. 



3rd. In many cases we are not sure if the dissolved electrolyte is to be regarded 

 as free from water or as a hydrate (in alcoholic solutions as an alcoholate). By the 

 researches of Graham, Riidorfl^, and others, we know that such hydrates exist in the 

 solutions. Now Professor F. Kohlrausch, in his already quoted memoir (p. 201), 

 has found no influence of the formation of hydrates. On the contrary, Professor 

 Eilhard Wiedemann, of Erlangen, has observed that there is a definite influence. 

 You know that chloride of copper in very concentrated solutions is green, in more 

 diluted blue. This change of colour is, without doubt, produced by the combina- 

 tion of the salt with the water. In the same manner diluted blue solutions grow 

 green by a rise of temperature. Now the electrical conductivity of solutions of this 

 salt at each rise of temperature of 5° C. augments nearly by the same amount up 

 to 60° 0. ; but above 60° 0. the change of conductivity for 5° falls off" in a remark- 

 able manner. The same is to be observed when the solution contains free hydro- 

 chloric acid. A solution of chloride of sodium — which salt forms no hydrates at 

 not too low temperatures — shows no such irregularities. Its conductivity rises 

 faster than the temperature up to higher temperatures. These experiments shall 

 be fiu-ther followed up. Without doubt the number of the salt molecules combined 

 with hydratic water must increase with the dilution. 



It seems to me at present impossible to avoid these disturbing circumstances, 

 which nevertheless should be put into consideration. Therefore we should not 

 content ourselves with the determination of the conductivities of very diluted 

 solutions, by which certain complications are avoided, but new ones may be intro- 

 duced. Only by the study of gradually changing concentrations of the solutions at 

 difffereut temperatures, and comparing their electrical behaviour with their other 

 physical properties, may we get an insight into these different conditions. 



The formation of hydrates of the salts and their dissociation must also be 

 understood before we can enter with any hope of success into further discussion of 

 the highly interesting and important question, principally treated by Dr. Svante 

 Arrhenius, whether the eventual formation of complex molecules and their dis- 

 sociation by further dilution may have an influence on the number of molecules 

 decomposed by the current, and on the resistance of the solution. 



For the same reason the question how much of the resistance may be due to the 

 friction between the imdecomposed salt and the dissolving medium, which have a 

 certain difference of electrical potential between each other, and therefore are trans- 

 ported in opposite directions by the current, must be postponed for further researches. 



IV. Another point which merits ample consideration and serious criticism is 

 the supposed relation between molecular conductivity and chemical composition. 

 The conductivities of dissolved salts give no such immediate relation ; nor is any to 

 be expected, since they are dependent upon the sum of the velocities of the two 

 ions. The same may be said about the conducti\'ities of worse conductors, the 

 organic compounds, &c. 



Great care should be taken in these researches to employ only chemically pure 

 substances. After the interesting observations of Mr. Ramsay about the boiling 

 points and vapour tensions of organic compounds, it seems not to be sufficient to 

 content ourselves with getting them from even the best chemical works and testing 

 them by their boUing points ; especially as the conductivities of bad conductors 



