Vol. 58. ] THE MINERAL ANALYSIS OF ROCKS. 163 
10. A Process for the Minerat Anatysis of Rocks. By Wintiam 
Jounson Sottas, D.Sc., LU.D., F.R.S., F.G.8., Professor of 
Geology in the University of Oxford. (Read January 22nd, 
1902.) 
Ir seems now to have become generally recognized that a natural 
classification of igneous rocks must rest ultimately on a chemical 
basis. The petrographer has thus become confronted with the 
necessity of an indefinite multiplication of chemical analyses, and 
may well shrink from the vast expenditure of time and labour that 
this involves. For many rocks there is no escape from this un- 
pleasing prospect, but in the case of others, that is, all holocrystalline 
kinds which are not too fine-grained, the question suggests itself 
whether a quantitative estimation of the mineral composition might 
not solve the problem with sufficient accuracy. Provided that the 
chemical composition of the mineral constituents does not vary 
capriciously, but is definitely related to their physical properties, 
such for instance as their specific gravity, there would seem to be 
no good reason why it should not; while, as I hope to show later by 
instances, there even appear to be grounds for concluding that the 
question can be more positively answered in the affirmative. 
In attacking the problem, the first step is to devise a method 
for obtaining a quantitative estimation of the mineral 
composition of a rock; and I may therefore proceed at once 
to describe a process which has been found to yield remarkably 
accurate results, 
A fragment of the rock required for analysis having been selected, 
it is reduced to powder, by the usual process of crushing in a steel 
mortar and sifting through copper wire-gauze. The meshes of the 
sieve should number on the average 12 to the square millimetre, with 
a variation not greater than 2 on each side of this, The whole of 
the fragment of rock should be powdered, and none of the grains 
rejected. 
The specific gravity of the several components must next be deter- 
mined, by means of a diffusion-column such as has frequently been 
described elsewhere. The heavy fluid used in this process should 
be methylene-iodide; acetylene-bromide is much less expensive, 
but its viscosity is far greater, and its vapour in a closed room may 
prove very injurious to the eyes. As is well known, the density 
of methylene-iodide may be increased up to 3°6 by saturating it 
with iodoform, or it may be diminished by dilution with carbon- 
tetrachloride, xylol, or benzol. It is convenient to have always 
ready for use a series of methylene-iodide and carbon-tetrachloride 
mixtures of definite specific gravity, differing from each other by 
intervals of 0°1, say from 3°34 to 2°44. 
A long, narrow, glass tube, graduated in centimetres, and fitted 
Q.J.G.8. No. 230. N 
