226 



REV. S. M. JOHNSTON 



curves are all made up of two portions, the one a rectilinear portion running from 

 higher to lower values of Mv/m* (and consequently from higher to lower dilution) 

 parallel to the ,u v /moo axis, and approximately along the 520 elevation constant line ; 

 the other a portion, either approximately rectilinear or but slightly curved, and running 

 away from the ^ v /m» axis and towards the elevating constant axis. The various curves 

 thus appear on the diagram as having coincident portions forming a common trunk and 

 non-coincident portions forming branches. 



It is not supposed that in every instance the curves change direction so sharply as 

 indicated by those drawn. 



From Professor * MacGregor's discussion of freezing-point data, it may readily be 



ki 



kj 



03 



ki 



3 4 5 6 7 



GRKEQS. PER LITRE = y 



9 kj 



Fig. 15. 



shown that a similar diagram with freezing-point depressions as abscissae would 

 give curves of the same kind, but that the curves would frequently be found to 

 branch off from the common trunk at points corresponding to higher dilution than 

 in the boiling-point diagram. This would seem to indicate what Biltz t has found, 

 that from the elevation of the boiling-point point of view salts act more normally 

 than from the depression of the freezing-point point of view. 



Passing on from these observations, we will now consider the meaning of the 

 high values which have been obtained for the elevation constant for concentrated 

 solutions. 



Can they be explained on the hypothesis that the ions and the molecules have 



* Phil. Mag., 1900, p. 505. + Z eit. phys. Chemie, 40, 204 (1902). 



