in the Analogy between Solutions and Gases. 99 



dissociation of which, under ordinary circumstances, does 

 not appear probable ; in aqueous solution, for example, the 

 majority of salts as well as the stronger acids and bases undergo 

 dissociation ; and hence Raoult did not discover the existence 

 of so-called normal molecular depression of freezing-point 

 and lowering of vapour-pressure until he investigated organic 

 compounds ; their behaviour is almost without exception 

 regular. For these reasons it may have appeared daring to 

 begin by giving prominence to Avogadro's law in its applica- 

 tion to solutious ; and I should have shrunk from this course 

 had not Arrhenius pointed out to me the probability that salts, 

 and analogous bodies, decompose on solution into their ions ; 

 in fact, substances which obey Avogadro's law are, as a rule, 

 non-conductors, suggesting that in such cases no dissociation 

 into ions occurs ; and further experimental proof exists for 

 other liquids, since by Arrhenius's assumption the deviations 

 from Avogadro's law are calculable from the conductivity. 



However this may stand, an attempt is made in the follow- 

 ing pages to take account of such deviations from Avogadro's 

 law, and, by help of the application of Boyle's and Gay- 

 Lussac's laws to solutions, to develop Guldberg's and Waage's 

 formula so far as is possible. 



The change in the expressions given above caused by these 

 deviations is easily sketched. 



The general expression for Boyle's, Gay-Lussac's, and 

 Avogadro's laws, shown on p. 90, is 



APV = 2T; 



and this changes, if pressure is z-tirnes that of this equation, 

 into 



APV=2«T. 



Hence, in a reversible cycle, the work will be ^-times that 

 previously done ; this alteration is easily applied to the former 

 statement of Guldberg's and Waage's formula. Recurring 

 to the final stage of the cycle described on p. 97, 



(3)+(6)=0, 



the work corresponding to (3) and (6), which wore formerly 



%'2a ti T -- 1 , and — %2a t T -p-^, is now increased {-times ; hence 



the equations become 



v / . d? u . dP\ n 



and on integration, 



X(aJ.. log P — ai log P,) = constant. 

 H2 



