1903.] ° RICHARDS—SOLVENT IN CRYSTALS. 83 
A much safer plan, where it is practicable, is to fuse the cystal- 
line precipitate. With the freedom of motion given by the liquid 
state volatile iinpurities usually soon escape, leaving the fused mass 
free from them. It is true that there is often danger that a portion 
of the substance itself will evaporate, or at least that it will attack 
the containing vessel ; minute precautions must be used to avoid 
these causes of error. In extreme exigency the electrolysis of the 
fused mass, a plan suggested by Richard Lorenz, may be used as a 
means of destroying the last traces of water; but in most cases 
these may be’ swept out by the vapors from easily decomposed 
ammonium salts, just as bubbling air sweeps out carbonic acid 
from its solution. 
A yet further and better thermal means of preparing a substance 
free from water is to vaporize and condense or sublime it in a per- 
fectly dry atmosphere. Here crystals form under conditions 
excluding water, and the danger is wholly overcome. Unfortu- 
nately other impurities are usually introduced from the walls of the 
vessel used for the sublimation ; but frequently these may be found 
by analysis much more directly and precisely than the water could 
be. 
A chemical method of disintegrating the structure of a substance 
crystallized from'a solution has been alluded to above. This 
method, although not always available, may sometimes serve when 
the other methods are inapplicable, and is often of great use in 
preparing chemically pure substances. This procedure, like the 
others, has been used in individual cases for years ; but it does not 
seem to have been emphasized as a general method. 
In brief the chemical method is as follows: The substance is 
crystallized from the solution, not directly in its desired form, but 
rather in chemical combination with a large quantity of some other 
substance which may be volatilized by suitable subsequent treat- 
ment. ‘The clear, dry crystals are then subjected to this decompos- 
ing treatment, and the volatile constituent is expelled. The 
substance sought is thus left in the form of a porous mass, a skeleton 
of the former crystal, in which every cell enclosing mother-liquor 
has been opened by the chemical disintegration of its walls. From 
such a skeleton soluble impurities may often be washed out by 
lixiviation, and volatile ones escape at once. 
For example, it is easier thoroughly to dry sodic carbonate when 
this salt is crystallized with its maximum amount of crystal-water, 
PROC, AMER. PHILOS. 80C. XLII. 172. C. PRINTED MAY 8, 1903. 
