732 JOHN JOHNSTON 



composition of a rock mass as we see it now is no absolute criterion 

 of what was the total composition of the magmatic solution from 

 which it separated; for we are ignorant of the character and 

 amount of the volatile components which were present in the 

 original liquid magma. 



In discussing metamorphic processes — or indeed in any discus- 

 sion of the effects producible by pressure — one must remember to 

 distinguish carefully between uniform (or pure hydrostatic) pres- 

 sure and non-uniform compression, which is equivalent to a stress ; 

 for the former is incompetent to produce some of the results which 

 may be brought about by the latter. This distinction is important; 

 failure to observe it has led to many apparently contradictory state- 

 ments, and consequently to some confusion. The main difference 

 is that whereas a homogeneous solid body recovers its original 

 dimensions and properties when relieved from a uniform pressure, 

 it is changed permanently by a stress, provided that the stress 

 exceeds the elastic limit of the solid; this residual effect is evident 

 as a change of shape — a deformation — which is accompanied by a 

 change in several other properties, e.g., in the case of metals by 

 changes in density, electric resistance, and thermoelectric power. 

 That the effects of stress on a crystalline aggregate outweigh those 

 of uniform pressure follows from thermodynamical principles if we 

 identify a deformation under stress with a real, partial, local tem- 

 porary melting, on the basis that the stress acts upon the solid but 

 not upon the liquid phase.^ This hypothesis, which is not in con- 

 flict with any direct experimental evidence, serves to correlate a 

 number of observations on the flow of rocks and of metals, and has 

 in recent years been utilized by several English metallographers 

 as a means of accounting for phenomena connected with the impor- 

 tant question of the hardening of metals. In this connection it may 

 be pointed out that the circumstance that a solid has been deformed 

 indicates, not that its strength was necessarily small, but that the 

 shearing forces to which it was exposed were great enough to over- 

 come the resistance to deformation offered by the solid; in other 



' This hypothesis, and some of its consequences, have been discussed in previous 

 papers; see Johnston and Adams, Am. Jour. Sci., XXXV (1913), 206; Johnston and 

 Niggli, Jour. GeoL, XXI (1913), 599. 



