■200 



EEV. S. M. JOHNSTON 



obtained above. The fact that in the case of each salt the curves run parallel, and 

 the difference in the elevation which they give for any given concentration is thus 

 independent of the concentration of the solution, makes it probable that the source 

 of error lay in the observation affecting all the elevations, i.e. that of the boiling point 

 of water. If I am correct in thinking that the above large values are due to error in the 



7 Z -3 -4- 5 -6 -7 



ELEVATION OF BOILING TEMPERATURE 



Fig. 2. 



determination of the boiling point of water, then in a series of determinations in which 

 increasing amounts of salt are added to the same water, the elevations observed would 

 all be affected by the same error, which would, in general, be different for different series 

 "I experiments. If so, and if series of observations be made for the same salt, and if 

 observed elevation be plotted against weight of salt added for several series, we might 

 e :pect to obtain parallel, but in general not coincident, curves. 



I" find out whether or not the determination of the boiling point of water is more 

 liable to he in error than that of a, solution, I applied heat successively of different 



