THE BEHAVIOR OF INCLUSIONS IN IGNEOUS MAGMAS 563 



back to the question of the effect of granitic magma on carbonate 

 inclusions. 



We have noted the silication of the inclusions and the consequent 

 supposed desilication of the liquid. In discussing investigated sys- 

 tems we have already found, however, that saturated liquids cannot 

 have ingredients subtracted from them at random without causing 

 precipitation of other ingredients so that the effect on the liquid is 

 not mere impoverishment in the ingredient subtracted. This may 

 perhaps be made clear by a simple example. If we had a solution 

 of salt at — 20° C, and any hypothetical substance was placed in the 

 solution that withdrew the salt from it, the result would not be sim- 

 ply the leaving behind of liquid water. The reason is simply 

 that liquid water cannot exist at — 20° C and the actual result 

 would be that, as each small amount of salt was removed, a small 

 amount of ice would form and, when all the salt was withdrawn, all 

 of the water would have become ice. For the maintenance of 

 liquidity, the salt and the water are necessary each to the other. 

 And so it must be with a saturated granitic solution. Remove the 

 silica from it and other substances must be precipitated. Now the 

 reaction of granitic magma with inclusions of carbonate rock is not 

 a simple addition of silica to the latter but usually other substances 

 are added as well, these being such as to convert the inclusions 

 ultimately into diopsidic pyroxene or hornblende. The reason for 

 the formation of these phases is that they belong at an earlier stage 

 of the reaction series than the biotite with which granitic magma 

 is normally saturated. The subtraction of the substances necessary 

 to produce these minerals must, for reasons outlined above, cause 

 concomitant precipitation of the other phases normally formed from 

 granitic magma, principally feldspar. Thus the action described 

 must bring about an exhaustion of the liquid by causing precipita- 

 tion. There seems to be no reason for believing that it could first 

 exhaust the free silica, leaving a feldspar-rich liquid, then, upon 

 further action, cause removal of some of the silica from the feldspar 

 liquid leaving a liquid containing feldspathoid molecules. 



A reaction of the kind described, that is, a using up of some silica 

 to form diopside with the consequent precipitation of feldspar and 

 quartz, would seem to me to be the mode of formation of the 



