44 ROYAL SOCIETY OF CANADA 
ture throughout both these readings. The tubes were now removed, the 
ice melted, and the same operation repeated for a second observation. 
The freezing-point of a hydrochloric acid solution of concentration 0:104 
gramme-molecule was found by this method to have the following values 
on the scale of a Beckmann thermometer : 
(1) 2°3086 (4) 2°3083 
(2) 2°3084 (5) 23084 Mean 2°3084 
(3) 2°3080 (6) 2°3086 
Thus results with a maximum variation of 0-0004 from the mean 
value could be obtained. The freezing-point of water at the time of the 
above observation was 2°6782 according to the same scale. Thus the 
depression of the freezing point of this solution, which is the difference 
between the freezing-points of water and the solution, is 0°3698 degree, 
As a change in the atmospheric pressure would cause a corresponding 
change in the thermometer, the freezing-point of the water used was 
determined every few hours. 
Since the freezing of my solutions was started about 0:1 degree 
below the freezing-point, the correction for the change in concentration, 
and therefore in the depression of the freezing-point, due to the formation 
of ice, is beyond the above limit of error of observation. The follow- 
ing method proposed by Raoult’ was used for determining it. The 
values of the lowering of the freezing-point of a solution are observed 
when different degrees of over-cooling are used. These observed 
depressions (C’) are then plotted on co-ordinate paper as ordinates 
against the over-coolings (S) as abscissæ, and it is found that the points 
lie practically on a straight line. The line joining these points is then 
produced and will cut the ordinate axis at a point which corresponds to 
the freezing-point depression of the solution when the over-cooling is 
zero. In other words we have obtained the true depression of the 
freezing-point (C). Thus this relation holds 
CEO LK CS a TONI es 
in which ©, C’ and S are known. The constant, A, can therefore be 
determined. With the described apparatus I found K to have about 
the value 0-016 for solutions of hydrochloric acid. Unfortunately data 
for this determination with sulphuric acid solutions were not obtained ; 
and I have assumed that Æ has in this case also the same value (0-016), 
since Raoult found that for two electrolytes and two non-electrolytes the 
values obtained differed very little from one another. The following 
results have thus been corrected by means of the above relation with this 
value of K. The correction thus obtained expressed as a percentage of 
the observed depression is about 0°16. 


1 Ztschr. f. phys. Chem., 27, 643, 1898. 
