1903.] RICHARDS—SOLVENT IN CRYSTALS. 35 
between true solution and colloidal solution in liquids; but for 
present practical purposes this difference need not be further 
emphasized. 
All these considerations have been carefully heeded in the recent 
determinations of atomic weights made in this laboratory. 
Since the fundamental properties of material have probably not 
changed since the archzan times of mineral growth, natural crystals 
must have been subject to the same effects as those grown to-day, 
and all except those formed by sublimation must contain traces of the 
solutions from which they once separated. In some cases, of 
course, the very slow formation might reduce the inclusion toa 
very small amount. ‘Those minerals coming from aqueous solutions 
would be supposed to contain accidentally enclosed water (often 
erroneously confounded with the true water of constitution), while 
those separating under fused or metamorphic conditions would con- 
tain non-volatile impurities taken from their immediate surroundings. 
Of these impurities the non-volatile ones must certainly remain ; 
and many of the volatile ones may also, if closely enough 
imprisoned. As a matter of fact, we are very familiar with the 
traces of impurity in natural crystals, even the clearest of diamonds 
leaving some ash on combustion. 
While the selective effects of adsorption and solid solution and 
the results of subsequent pressure must complicate the interpreta- 
tion of these facts, and forbid their immediate quantitative utiliza- 
tion, it seems to me nevertheless that these enclosed traces of 
impurity might be used more often than they are used in geological 
reasoning, in order to discover the media from which crystals have 
separated, and hence the mechanism of their formation. Thus a 
phenomenon very troublesome to the chemist might become very 
useful to the geophysicist. 
The contents of this paper may be summarized briefly as follows : 
(a) Experiments are recorded and quoted showing that many, if 
not all, crystals separated from solutions contain included mother- 
liquor. » 
(2) The experiments show also that before the mother-liquor can 
be eliminated by pulverization the adsorption of water from a moist 
atmosphere begins to augment the weight of the substance. 
(c) It is pointed out that this adsorption cannot be wholly over- 
come in the case of hydrated salts without a loss of water of crys- 
tallization also. Hence hydrated salts cannot be accurately weighed 
according to any usual procedure. 
