VOLUME CHANGES IN METAMORPHISM 553 
As the matrix is usually made up of alternating lamellae of mica 
and quartz it is possible that the quartz is dissolved in the direction 
of stress and deposited perpendicularly to it, while the mica Is stable 
under prevailing conditions. This explanation might apply to 
some cases where the garnets have the appearance of having grown 
by forcing the schist apart. Anyway, such an action would involve 
an increase in volume while certainly the whole tendency in crystal- 
line schists is more toward compression than expansion. The 
microscopic evidence, therefore, favors the view that replacement 
in crystalline schists has taken place by equal volumes, though the 
possibility cannot be denied that under heaviest stress the mode 
of replacement may be so altered that a smaller volume results. 
The occurrence in many of these rocks of abundant crystals 
of heavy aluminum silicates raises the question whence came this 
concentration. The alumina could not have been derived only 
from the replaced host minerals; it must have been concentrated 
from material some distance away. It seems then that it would be 
necessary to assume the presence of moving liquid or gaseous solu- 
tions. And if moving solutions are admitted there is no essential 
difference between replacement in this case and in that of static 
metamorphism. These solutions would have carried other material 
from the outside, and the difference in volume between the newly 
formed heavy minerals and the dissolved host minerals would be 
equalized by new deposits whereby the demand for replacement by 
equal volume could be satisfied. 
It would seem then that the conception of metamorphism under 
stress here outlined involves somewhat free circulation and limited 
addition and subtraction of material, with strong tendency toward 
the preservation of volume. 
This is to some degree in opposition to the views of Rosenbusch 
and Grubenmann but in line with a recent paper by Leith and 
Mead,' in which the convergence to mineral type in dynamic meta- 
morphism is emphasized. Duparc and others have shown us that 
the conversion of hornblende to uralite is not a simple paramorphism 
but a metasomatic process. Leith and Mead point out that the 
1C. K. Leith and W. J. Mead, ‘‘Metamorphic Studies,” Jour. Geol., XXIII 
(1915), 600-607. 
