4IO N. L. BOWEN 



the crystals more closely packed. There is an intermediate neutral 

 zone in which the sheet retains its original width and the crystal 

 mesh remains undisturbed. This portion therefore acts as a 

 permeable partition through which the interstitial liquid of the 

 thinned portion passes.^ As the walls begin to move apart in the 

 thickened portion the crystal mesh will part along a median plane, 

 for there it is weakest, and the interstitial liquid from the thinned 

 portion will flow in to occupy the space afforded. At no time will 

 there be any space in the sense of empty space. The movement of 

 liquid into the space will be absolutely concurrent with, and 

 perhaps a contributing cause of, the development of the space. 



The action will continue until a proper 

 ^ ;-y- v':n.S-;y:;';.-- V:; ?::.■;.: y:'^'-;:i^v.v V balance of the forces involved is reached. 

 ^ B In certain sections the thickness of the 



sheet may be increased to several 

 times its original value, and the late 

 differentiate may be a very thick cen- 



FiG. 3. — Sections of a tabular ,ii j-j ^ ^ .• ^^ • ^ 



, , -11 / \ 4. tral band, mdeed several times as thick 



mass, dyke, or sill: {a) at an ' 



advanced stage of crystalliza- as the narrow bands of the early differ- 

 tion; ih) after warping of walls entiate on either wall. The thinned 

 at this stage and squeezing of ti^^s will normally be thinned only 



residual liquid (late differentiate) 



into expanded section. a moderate amount, so that relatively 



large areas must be thinned. The late 

 differentiate would again have an intrusive relation to the earlier 

 material on the walls, with, however, ''welded" contacts and locally 

 a transitional relation of a more or less abrupt nature. This action 

 could take place only at a late stage of crystallization, when the 

 crystal mesh had very considerable strength, if only for the reason 

 that the neutral zone must have sufficient strength to withstand a 

 considerable difference of hydrostatic pressure on opposite sides of 

 it, for there would be very great frictional resistance to the flow of 

 liquid. 



' For the purpose of presentation of the conception in particularly simple form 

 this definite neutral zone is assumed. Its existence is, however, not essential to the 

 action described. It is merely necessary that the breakdown of the crystal mesh 

 should go on progressively as stated on p. 408. This will bring it about that the por- 

 tion not yet broken down acts at any instant as a permeable partition preventing the 

 passage of crystals and permitting the passage of liquid from the portion breaking 

 down at that instant. 



