[MACGREGOR] DEPRESSION OF THE FREEZING-POINT 9 
determine the depression itself within less than 0:0002 degree. An error 
of this magnitude would in the case of the most dilute of the above 
solutions amount to about ! per cent, and in the case of the stronger, to 
from 31, to 1 per cent. Inthe values of the equivalent depressions, it 
would amount to about 4 or 5 units in the fourth significant figure 
for the weakest solutions and to about 1 or 2 for the stronger. Young 
experimenters with much less satisfactory apparatus could not be 
expected to attain such accuracy as Raoult, and we may safely assume 
the limit of error of the above observations, apart from the accidental 
slips which have been roughly corrected, to be 7 or 8 in the fourth 
Significant figure for the weakest solutions, and at least 3 or 4 for the 
stronger. 
Assuming some such value for the limit of error, it is obvious that 
for the more dilute solutions of most of the above series the differences 
between calculated and observed values are within the limit. In other 
words, the curves obtained by plotting equivalent depression against 
ionization coefficient, are practically straight lines at sufficiently great 
dilution, though they exhibit marked curvature in one sense or the other 
and of greater or smaller rate, as the dilution diminishes. 
The hydrochloric acid curve exhibits more rapid curvature than the 
others, and can hardly be said to be practically straight even at the 
greatest dilution examined, though it is clearly tending to become 
rectilinear. Of the two series of observations on this electrolyte, both of 
which show the same marked curvature, the first was made by Mr. 
Barnes at an early stage, and the second at a later stage, of his experience 
of observations of this kind. The second is thus the more trustworthy of 
the two. 
The depression observations of the sodium chloride solutions were 
made at the same early stage and are somewhat defective, but when 
corrected by the aid of Loomis’s results, the curve for these solutions is 
seen to be practically a straight line up to a concentration of 0-1. 
The potassium chloride curve, whether judged by the uncorrected or 
the corrected depression observations, is practically straight up to a 
concentration of 0-5. 
The potassium sulphate curve begins to exhibit rapid curvature at a 
concentration of about 0°1, the sodium sulphate curve a less rapid 
curvature at 0-3. 
The sulphuric acid curve shows little curvature at any part of the 
concentration range examined. 
