Components of an Igneous Rock — Williams 39 
would have a definite position which could not be varied. 
If, however, there happen to be a considerable range be- 
tween whose limits the pure minerals will respectively rise 
and sink, in other words, between the specific gravities of 
the two minerals in question, the density of the solution 
may be adjusted to any position within these limits and yet 
give a perfect separation if the minerals are pure and entirely 
apart from each other. That they are never entirely freed 
from one another in breaking up a crystalline rock need not 
be emphasized. There are always particles of one clinging to 
every other and it is usually impossible to say whether or 
not more of one clings to a second than of the second to the 
first. By moving the density of the solution therefore nearer 
that of one mineral than of another to be separ£.ted from it, 
more or fewer of the particles composed of parts of each 
will be shifted one way or the other with consequent modi- 
fication of results. 
The only feasible plan that appears to avoid some of 
the discrepancies that are apt to arise in this connection is 
to first estimate approximately the relative proportions of 
the minerals by means of the microscope, or, the heavy solu- 
tion itself. Then, assuming, unless accurate data can be ob- 
tained, that equal portions of the minerals cling to each 
other, adjust the specific gravity of the heavy solution to a 
position between the extremes determined by the relative 
proportions of the two minerals in the rock. 
This same factor becomes of increasing importance in 
the separation of minerals whose specific gravities are very 
nearly alike if by any outside influence, the heat of the body, 
or a draught of air, the density of the solution is altered ever 
so slightly during a separation. Precautions are necessary 
to maintain surrounding conditions constant. 
The facts just pointed out militate against the scientific 
accuracy of the heavy solution method for making quantita- 
tive separations of mixtures of minerals. Rosenbusch sums 
up the reasons why the method cannot be accurate in the 
following:* 
I. The impossibility of preparing a powder which shall 
consist of nothing but homogeneous grains ; 
*lDDiNG9' Translation, 1900, p. 107. G erman Edition, p. 249. 
