BE RICHARDS—SOLVENT IN CRYSTALS. [April 4, 
. than it is when the salt is crystallized immediately in the form 
devoid of crystal-water. Again, it is far easier to obtain pure iron 
by the reduction of the oxide in hydrogen than it is by electrolytic 
precipitation direct from a solution. 
Yet another chemical method might sometimes be helpful, 
although inherently of a somewhat restricted usefulness. In a 
careful study of the behavior of gases included in oxides,’ it was 
found that in these compounds oxygen could work its way out from 
a cell in which it exists under pressure, while nitrogen could not. 
It is possible that this fact is typical of a general principle; that 
any substance may escape from. pressure when this substance forms 
one of the easily dissociated components of the containing 
walls, by a process of alternate dissociation and recombination of 
the materials constituting these walls. Thus certain hydrated 
salts, heated to a temperature of perhaps 120° in superheated steam 
at atmospheric pressure, might in time part with their enclosed 
water without losing their water of crystallization. This inference 
will be further tested in the near future. 
In this connection, one other point must be strongly emphasized, 
because of its important bearing both upon chemical purification 
and upon dynamic geology. The microscopic cells in a crystal 
contain not only the solvent, but also all the other substances in 
the solution. They are miniature samples of the solution, not per- 
haps in strict quantitative measure, because of varying adsorption, 
but at least qualitatively. 
Hence, if a chemically pure substance is desired, great pains 
must be taken to keep away from the solution anything which can- 
not be expelled or extracted from the disintegrated crystal, after 
the enclosures have been opened. For example, pure iron must be 
obtained by the reduction of oxide or hydroxide obtained from the 
nitrate, not from the sulphate. If the latter salt were used, the 
iron would probably contain sulphur. This practice of continually 
avoiding impurities obviates also the possible danger from ‘‘ solid 
solution,’’ as van’t Hoff pertinently terms the homogeneous distribu- 
tion of foreign matter in a solid. Solid solution, or occlusion, may 
be said to be the limiting case of cuc/usion. In this limiting case 
the enclosing cells are so small as to contain only single molecules. 
The distinction between occlusion and inclusion in solids is 
theoretically interesting, corresponding perhaps to the difference 
1 Richards, dm, Chem. Four., 20, p. 701 (1898). 
