1045 
of the concentration, yields for NaCl and KCl almost straight lines 
whereas those for MgCl, and CaCl, exhibit a slight bending upwards 
at the greater concentrations. 
The values of 7 exhibit with MeCl, and CaCl, a minimum for 
c= + 0,2,. whilst also here, contrary to the calculation from A, 
those of MgCl, are larger than those of CaCl,. Also the difference 
between NaCl and KCl is here smaller than followed from the A 
determinations. 
The most trustworthy observations of other investigators mostly 
agree well with those of mine for KCl and NaCl, whereas those 
for MeCl, and CaCl, exhibit at both sides deviations of at most 
2°/, of the value of ¢ according to the graphic representation. 
The determinations of the boiling points were carried out in metal 
vessels internaliy silver plated and surrounded by a steam jacket. 
In order to avoid corrections for the barometric pressure, a second 
vessel with pure water was always boiled under exactly the same 
conditions as the one containing the saline solution. An exchange 
of the thermometers in the two vessels could always take place without 
any danger of escape of vapour by placing these thermometers ina 
thin-walled tube containing mercury. These tubes were placed in 
the two vessels at an equal depth so that no correction for the hydro- 
statie pressure was required. By lengthening the refrigerating tube 
until it penetrates lower into the boiling vessel it was avoided that 
the colder reflux water had an influence on the thermometer bulb. 
A retardation of boiling was counteracted by placing in each 
vessel 200 grams of clear glass beads and 10 silver tetrahedrons. 
A correction of the concentration for the water withdrawn from the 
solution by evaporation was applied. By experimenting this was 
determined as 0.2 °/, of the concentration when the vessel contained 
250 ec. of liquid. The concentration was determined before as well 
as after the boiling and then yielded no measurable differences. The 
necessary scale corrections were introduced on the thermometer. The 
calculation of c’ from c was executed as directed above. Here also, these 
quantities relate to a temperature of 15°. 
The graphic representation in which ¢ was again plotted as 
function of ec shows that the faintly bent curved lines for NaCl and 
KCl almost coincide, NaCl now being situated a little higher than 
KCl. Just as with the solidifying points, MgCl, is again higher than CaCl,,. 
In the caleulation of z¢ the theoretical value 0,52° was again chosen 
for the molecular increase of the boiling point in 1000 ce. of water. 
The values of # now exhibit with all salts a minimum, with NaC! 
and KCl for c’ =+0,3 and with MgCl, and CaCl, for c’ = + 0,2. 
69* 
