38 Professor A. Crum Brown [Jan. 31, 



is unity in the cases where observation by Raoult's method gives results 

 agreeing with the formula PV = RT, in other cases it is greater or 

 less than unity, and indicates the extent of the disagreement. Arrhe- 

 nius, to whom Van't Hoff showed these numbers, pointed out that all 

 the substances which had i greater than unity were electrolytes, and 

 that the deviation had to do with their splitting up into ions. 

 Arrhenius * had before this time (1887) been working at the subject 

 of electrolysis and of the relation between the readiness with which 

 substances undergo electrolysis and the readiness with which they 

 enter into chemical reactions. He had been looking for an explana- 

 tion of the fact that the conductivity of a solution of an electrolyte 

 is not proportional to its concentration, and had come to the conclusion 

 that this must depend on some of the molecules of the solute being 

 " active," that is, taking part in the conduction, while others were in- 

 active, behaving like molecules of a non-electrolyte, and that the pro- 

 portion of active molecules increases with dilution. 



Van't Hoff's factor i enabled Arrhenius to give precision to these 

 ideas, and in 18871 he formulated the theory that the " active " mole- 

 cules were those which were split into ions. It was now possible 

 to calculate i in two ways and compare the results. Arrhenius gives 

 a list of eighty-four substances, for which there existed at that time 

 data for such calculations, and calculating the value of i as deduced 

 on his new theory from the conductivity, compares it with the value 

 of i derived from freezing-point observations in each of the eighty- 

 four substances. The agreement does not at first sight strike one as 

 very close, but there are several circumstances which have to be 

 considered in judging them. The whole mass of published observa- 

 tions was taken, the limits of probable error are very different in 

 different cases, and the freezing-point measurements were all made 

 at temperatures a little below 0°, while the conductivity measure- 

 ments were made at 18°. The comparison was made, not as a 

 demonstration of the theory, but rather as a preliminary trial with 

 such materials as were at hand. The real testing of the theory 

 necessarily came later. So I think we may agree with Arrhenius 

 that, considering all the circumstances, the agreement is not unsatis- 

 factory, except in the case of nine of the substances, and that most 

 of these nine cases are liable to suspicion on other grounds. In 

 1887, almost at the time when Arrhenius published the paper of 

 which I have just been speaking, Planck| discussed the subjects of 

 the diminution of the vapour pressure and the lowering of the 

 freezing-point in dilute salt solutions from the thermo-dynamic point 

 of view, and starting from the principle of the increase of entropy, 

 deduced formulae connecting these quantities with the molecular 



* Arrhenius, Bihang till kongl. Svenska vetensk. Akad. Handlingar, 1884 

 Nos. 13 and 14. 



t Arrhenius, Zeitscbrift f. physik. Chemie, i. p 631 (1887). 

 X Planck, Wied. xxxii. p. 495 (1887). 



