956 Proceedings of Royal Society of Edinburgh. [sess. 
Several series of experiments were performed on each of a 
number of salts to find the relation between the boiling point 
elevation of different series on the same salt. It was seen from 
plotting elevations of boiling temperature against weight of salt 
added, and drawing in the curves for different series on the same 
salt, that the curves obtained were approximately straight lines, 
and parallel to each other. Specimens of these are given for 
potassium nitrate, barium chloride, and lithium chloride on page 
957 ; ammonium sulphate and potassium chloride on page 958 ; 
and sodium nitrate on page 959. The straightness of the curves 
in the vicinity of the origin was also tested in some cases, and it 
was found that, so far as experiment could he carried, the curves 
continued approximately straight lines. 
The boiling temperature of the solvent under these considera- 
tions being looked upon as at fault, that is, its experimental value 
varying by one or two hundredths of a degree, and the boiling 
temperature of a solution being steady, it became desirable, in 
determining the boiling point constant, to use only elevations above 
the boiling point of a solution. I consequently computed, not the 
elevation of the boiling point per gramme particle produced by 
adding a given number of grammes of salt to one gramme of 
solvent, hut the elevation per gramme particle produced by adding 
given numbers of gramme molecules of salt to a solution, the 
expression employed being 
n m.W.AE 
O “ - - j 
(1 + n - la)A w 
where m, W, n and a are as before, and AE and A w increments of 
elevation of boiling temperature and grammes of salt added 
respectively. According to theory, the values obtained should he 
the same as those from the first formula if the E could he 
AE 
accurately determined. The ratio was found as follows : — 
J A w 
Let OW and OE he axes of weight of salt added and elevation 
of boiling temperature respectively. Let LR, the curve for any 
series of observations, found as seen above to he approximately 
a straight line, be produced to meet the axes in P. Through 
any near points Q and R on this curve draw parallels to the 
axes. 
