200 REPORT—1890. 
p. 347), ‘ This is not generally true. First, we have certain bodies which 
seem not to be decomposed by the current, though they exchange their 
elements with those of other compounds which are electrolytes. Take, 
for instance, anhydrous hydrochloric acid. It does not conduct. Never- 
theless, as Dr. Gore has shown, if you put it upon carbonate of lime the 
carbonic acid is chased away and chloride of calcium is formed. And, 
to give another example, the chloride of propyle is a non-conductor; 
nevertheless, when you treat it with bromide or iodide of silver the 
chloride gets changed into bromide or iodide. With just reason you 
may object that this is no proof, for perhaps the chloride of propyle is 
only a very bad conductor, therefore the current does not pass in a sensible 
way and we cannot observe the decomposition. In this respect we may 
refer to the researches of Mr. Bleekrode in Holland, and Mr. Bartoli in 
ltaly. 
Bat, on the other side, we find well-known electrolytes exchanging 
their ions with elements of other compounds which, without. any doubt, 
are not their ions. So, for instance, chlor-acetic acid (CH,CICOOH) or 
the ethylic ether of this acid, and iodide of potassium exchange between 
each other the chlorine and iodine, though assuredly the ions of chlor- 
acetic acid are not Cl and CH,COOH, but CH,CICOO and H.’ (See 
also Wiedemann, ‘Elec.’ vol. 2, p. 926, and Lodge, ‘B.A. Report,’ 1885.) 
If we adopt the dissociation hypothesis we may say that an electrolyte 
is a substance part of which is in a state of dissociation, each dissociated 
molecule being resolved into two parts, which form the ions in electrolysis. 
It remains to be considered whether there is any means of finding out 
(otherwise than by conductivity) whether there is any such dissociation. 
The processes of chemical reaction are, however, brought by the 
dissociation theory into close connection with electrolytic action, so that 
Hittorf’s classification can only be distinguished from the definition based 
on dissociation by the consideration that the Jatter goes a step further 
and explains and accounts for Hittorf’s empirical generalisation. The 
case of chlor-acetic acid and others similar are considered by Ostwald, and 
cause him to extend the dissociation hypothesis in order to include them 
(see below, p. 220). 
There remains, therefore, the definition forming the fundamental 
hypothesis of the dissociation theory, viz., that an electrolyte is a substance 
which contains some compound in a state of partial or complete dissocia- 
tion. It is upon this hypothesis that a great deal of recent work in 
electrolysis has been based, and nearly all the observed phenomena of 
electrolysis have been deduced from it. What the precise nature of the 
dissociation is may not be clear. The provisional hypothesis regards the 
dissociation of a compound in an aqueous solution as the resolution of 
the molecules of the compound into atoms or their chemical representa- 
tives which form the ions in electrolysis. Large strides have been made 
towards the formation of a mechanical theory of the electrolysis of solu- 
tions on this basis, some account of which will be given below. What 
is still wanting for the completion of the theory, besides the explanation 
of small numerical differences between calculated and observed results 
and the development of its extension to include the exceptional cases 
mentioned above, is the investigation of the mode in which the solvent 
acts in producing the necessary dissociation without itself being appre- 
ciably resolved. That may no doubt be forthcoming when the actions of 
different solvents have been observed ; in the meantime it is interesting 
