ON ELECTBOLTSIS IN ITS PHYSICAL AND CHEMICAL BEARINGS. 353 



state, as Faraday lias already remarked. Because of the regularities which 

 characterise dilute solutions, the dissociation hypothesis has reference almost ex- 

 clusively to matter in this state, and has, in my opinion, very successfully explained 

 a great many of its properties. 



3. Professor Armstrong asserts that solutions in absolute alcohol oppose a 

 practically infinite resistance. This view is not supported by the recent investi- 

 gations of Fitzpatrick and Hartwig, from which it seems to follow that ' any 

 neutral solvent renders electrolysis possible,' in accordance with the demands of 

 the dissociation electrolytic hypothesis. 



4. According to Professor Armstrong, on the dissociation hypothesis, both 

 water and salt are to be regarded as dissociated ; nevertheless, the dissociation 

 hypothesis assumes that conduction takes place only through the agency of the 

 atoms of the dissociated salt, and not at all through that of the water. From 

 chemical facts, e.g., the partial decomposition of certain salts into base and acid by 

 large quantities of water, it is very probable that water is to a very slight degree 

 dissociated into H and OH.' But experiments show also that this occurs to an 

 extent so extremely small that it will perhaps for ever be impossible to detect it 

 by measuring the conductivity of water. Also, I have essayed to show that the 

 only fact cited in favour of a sensible conductivity of water can be satisfactorily 

 explained without this assumption. In the same manner, different electrolytes, 

 e.g., HCl and OjHjOoH, are very differentlj' electrolytically dissociated by the 

 same quantity of water, so that the one conducts very well, the other hardly 

 sensibly ; but we know representatives of all intermediate stages between these two 

 extremes, so that there is no ground for assuming anything but a quantitative 

 difference. Experiments on the conductivity of pure water, or of dilute salt 

 solutions, lead to the assumption that water is not dissociated to any perceptible 

 degree. If, therefore, as Professor Armstrong asserts, the opinion is common 

 among chemists that in the case of the electrolysation of ammoniacal solutions 

 •water is (sensibly) electrolysed, then is this opinion a remnant of old views now 

 abandoned through force of incontestible arguments ? 



6. In his article Professor Armstrong makes no objection to the dissociation 

 electrolytic hypothesis, but says : ' I do not mean for one moment to assert that 

 anything which we know of the conditions on which chemical change depends 

 negatives beyond question the dissociation hypothesis.' 



7. Professor Armstrong interprets the figures of Lenz concerning the parallelism 

 between difl^usivity and conductivity of alcoholic aqueous solutions as an ' almost 

 conclusive argument in favour of his theoiw.' It must be remarked that there are 

 many exceptions to this rule. Thus, e.g., MgS04 and cane sugar diff'use nearly at 

 the same rate (Graham), although in aqueous solutions MgSO^ is a good con- 

 ductor, while cane sugar seems not to conduct at all ; consequently, the two 

 phenomena are not of the same order, as they should be according to Professor 

 Armstrong's hypothesis. ' The diminution in conductivity and also in dift'usivity 

 as the amount of alcohol is increased is most striking. If the solvent be neutral, 

 the substitution of alcohol for water should have little influence ; but if, as I 

 suppose, the solvent be active, alcohol being far less active than water, the effect 

 to be expected is precisely of the nature of that observed.' In coming to this con- 

 clusion a very important thing is lose sight of. The conductivity depends on two 

 circumstances — viz., the degree of electrolytic dissociation, and the friction which 

 the ions must overcome in their passage through the solution. The first circum- 

 stance must, as Professor Armstrong asserts, have but little altered in Lenz's 

 experiments with diluted KI and Nal ; the diminution in conductivity must conse- 

 quently have chiefly depended on the increase of the friction with the quantity of 

 alcohol, which is a priori very probable. Evidently one must expect that the 

 diffusivity also diminishes as the internal friction increases, and in this manner the 

 parallelism between conductivity and diffusivit)' in Lenz's figures is quite explicable 

 by means of the electrolytic dissociation hypothesis. 



8. ' Why, then, should it (the electromotive force") not give direction to the 

 moving molecules, if these are still possessed of "residual affinity " — i.e., if some 



' Wied, A7inalen, sxxiii., p. 58, and PJtil. Mag. (5), xxiv., p. 377. 

 1888. A A 



