40 Professor A. Crum Brown [Jan. 31, 



tions of the same electrolyte of different concentration are in contact 

 there is a difference of electric potential between them, but they 

 cannot be separated to any weighable extent in this way. In order 

 to separate from one another two gases uniformly mixed, a certain 

 calculable amount of work has to be done, so that after a gas has been 

 dissociated and wholly or partially converted into a mixture of the two 

 gaseous products, some work has still to be done to get them sepa- 

 rately. So it is also in the case of electrolytic dissociation ; but while 

 in the former case the decomposition work is the main thing, and the 

 separation work very small, in the latter it is quite the other way. 

 Here the heat of dissociation, that is the work spent in decomposing 

 the electrolyte into its ions, is small (indeed sometimes negative), 

 while the work to be done to separate the ions is always very much 

 greater. Indeed we may quite correctly say that in most highly dis- 

 sociated solution of hydrochloric acid the hydrogen and the chlorine 

 are still very firmly united, not indeed atom to atom, but each atom 

 of the one kind to all the atoms of the other kind within a certain dis- 

 tance from it. A man does not lose his money when he takes it out 

 of his pocket and puts it into a bank. He does indeed lose his rela- 

 tion to the individual gold and silver coins, and does not know and 

 does not care where these particular pieces of metal are, but he is 

 interested in knowing that they or their like are at his command, and 

 the same sort of work will be required to impoverish him whether his 

 money is in the bank or in his pocket. (I assume, of course, that the 

 bank of our present imagination cannot become insolvent.) 



I have said that the test of the theory would come later. It has 

 been going on since 1887, and if time would allow I could give you 

 many cases in which deductions from the theory have been found to 

 agree with close quantitative accuracy with experimental observa- 

 tions. I shall mention only the first, still among the most impor- 

 tant, namely Ostwald's determination of the affinity constants, and 

 his application of Guldberg and Waage's principle to the ions. I 

 could also give you instances in which there have been discrepancies, 

 or apparent discrepancies, and show how in some of these cases the 

 difficulties have been cleared up. The history of this theory has in 

 fact so far been that of every useful theory, for it is in this way only 

 that a theory does its work. I shall select two points for illustration, 

 not because they are more important than others, but because I can 

 illustrate them by means of experiments which do not occupy much 

 time, and can be made visible in a large room. The first has refer- 

 ence to the question, What are the ions in the case of a dibasic 

 acid ? As HNO3 gives as its ions H and NO3 so we might expect 

 H2SO4 to give 2H and SO4. But we find that until the dilution has 

 advanced to a considerable extent the ions of sulphuric acid are 

 maialy H and HSO4. This is quite in harmony with the chemical 

 action of H2SO4, for, as every chemist knows, at moderate temperatures 

 we have the action H2SO4 -h NaCl = HCl + NaHSO^ and the tem- 

 perature has to be raised in order to get the action NaHS04 -1- NaCl 



