as before, and as soon as the thermometer has become stationary the tube a is taken away (a and a' being at the same time closed with small corks) and after drying is again weighed correct to centigrams. The weight of a and sub- stance subtracted from the weight of a and solution gives the weight of the solvent. Walker and Lumsden 1 have modified this method, so that the volume, and not the weight of the solvent, is observed. Landsberger* s method has the advantage of greater rapidity over Beckmann's. The thermometer employed must be very sensitive and graduated in twentieths of a degree. Landsberger re- commends the use of a series of thermometers, each of which has a scale about 20 cm. long and registering only about ten degrees.
Bemark. — For the application of- these methods, the substance must not distil with the vapour of the solvent. Its boiling point must, therefore, be higher, by at least 130 degrees, than that of the solvent.
Tlworetical Deductions. — The two methods lead to the same result. They both show us that a solution whose grams of solvent boils t° higher than the solvent. Accepting the principles which are given at the beginning of this chapter, we may reason as follows : If in the 100 grams of solvent we had put p times less substance, that is, if we had made a 1 percent, solution, the rise in the boiling point would
have been - degrees ; and if we had dissolved one gram- P
molecule the elevation would have been - m degrees (where
P m is the molecular weight). We thus calculate the
molecular elevation of the boiling point, e, which is indepen- dent of the nature of the dissolved substance, and varies only with the solvent employed.
For any given solvent we can ascertain e by two means :
Wourn. Chem. Soc, 73, 502 (1898).— Tr.
�� �
