( 638 ) 
mined. In the case of KCl, 7 also reaches a minimum between the 
concentrations 0.1 and 0.5 gr. mol. From 0.5 gr. mol. upwards a 
rise of 7 is observed as the concentration increases. Above 0.5 gr. 
mol. the results are therefore, qualitatively, the same as with the 
micromanometer. Below this concentration 7 seems to increase again 
with the dilution. In the case of K NO; @ increases proportionally 
with the dilution just as I bave observed with the micromanometer. 
Both at 100° and at 0° a solution of K NO, seems to behave 
differently to solutions of K Cl and Na Cl. 
As regards the accuracy of the results, I may say that the greatest 
error of each thermometer was 0,002°. If these errors of the two 
thermometers have opposite signs, the error of observation amounts 
to 0,004°. As, however, each determination lasted 15 minutes, a 
reading being taken every 5 minutes, the average error must have been 
really less than 0,004° which is confirmed by the thorough agreement 
of the results obtained in the three series of observations on the 
Na Cl solutions. 
Earlier observations. 
W. LANDSBERGER!) has already found in 1898 that more con- 
centrated solutions of Na Cl show a rise in the molecular increase 
of the boiling point when the concentration increases. From his 
molecular weight determinations, I have calculated the following 
values for the molecular increase of the boiling point and for @. 
Nia CA. 
HR 
Concentration in Elevation of the Molecular perp ; 
gr. mols per 1000 grs. of H,O.| boiling point. weight found. boiling point. 
0.7145 0.676 32.2 9.46 1.82 
1.0581 1.026 31.4 9.70 1.87 
1.0872 1.080 30.7 9.93 1.91 
1.1077 1.067 31.6 9.63 1.85 
1.2427 1.235 30.6 9.94 1.91 
2.0735 pl Wear 29.0 10.50 2.02 
2.0855 2.186 29.0 10.50 2.20 
1) Zeitschr. f, Anorg. Chem. XVII 452. (1898). 
