Joty—Pressure on Separation of Silicates in Igneous Rocks. 879 
will be positive in sign—a change of increase—and will thus, 
under the influence of pressure, act in opposition to the negative 
change of volume, probably occurring on crystallization. We are 
not, however, swe that the volume-change of withdrawal from 
solution will be positive. As the withdrawal from solution and 
assumption of the crystalline form are accomplished simultaneously 
the algebraic sum of the volume-changes is what concerns the 
final result. The change in volume attending solution, or with- 
drawal from solution, is often very considerable in known cases. 
We are ignorant of what its amount may be in the case of silicates 
dissolving in a siliclous magma at temperatures near their melt- 
ing points. Hxperimental investigation presents considerable 
difficulties. 
But this is not all. We are hardly justified in assuming 
without question that the silicate molecule is first formed upon its 
crystallization. That in short the crystal is built up by successive 
additions of the more stable sub-molecules, such as CaO, A1,O,, 
Si0,, adding themselves independently to the crystal. We are, 
indeed, prima facie entitled to assume that molecules of the 
particular silicate were first formed in the magma, or were so in 
some cases. If this is the first stage in the order of events, the 
pressure influence must be considered here also. ‘The formation 
of such combinations as possess small molecular volumes would be 
favoured. Pressure might decide between two contending com- 
binations, conferring greater stability upon the one, and hence 
causing it to grow at the expense of, or to the exclusion of, 
another. 
It must also be borne in mind that, according to Maxwell and 
others, at every stage of temperature the viscous silicate-glasses 
would contain macro-molecules of various magnitudes, accounting 
for their partly solid, partly liquid state. 
Every petrologist has had the extreme complexity of the 
phenomena of rock formation forced upon him with the extension 
of his studies. I would here enter a word of warning against 
applying without very full consideration a formula, applicable to 
reversible processes only, to the very complex conditions (chemical 
and physical) attending volume change on the formation of silicates 
in igneous rocks. The extreme uncertainty introduced into the 
computation by indefinite melting points, slow changes of volume, 
