lxxii PROCEEDINGS OF THE GEOLOGICAL SOCIETY. [vol. lxxiv,. 



Here again the semi- calcareous sediments afford an instructive 

 contrast, for it is not uncommon to find them converted to a 

 coarsely crystalline aggregate of garnet or of diopside in contact 

 even with a dyke of moderate width. Such rocks, until they 

 have been Avholly reconstituted, have the special peculiarity of 

 providing their own solvent in the form of carbon dioxide set 

 free in the reactions, a property obviously conducive to that 

 rapidity of metamorphism on which I have already remarked. 

 It is further of importance that- the loss of this constituent implies 

 a shrinkage of the solid rock, which may often amount, according 

 to Barrell's calculation, to 30 or 40 per cent, of the volume. 

 This estimate, of course, ignores the pressure to which the rock 

 may be subjected during the process ; but it is clear that the 

 effect must be to enhance its permeability in the first place by 

 the gaseous solvent and later, when cooling has sufficiently 

 progressed, by liquid solutions. In illustration may be cited the 

 ' garnet -contact -zones ' which have acquired some importance 

 in recent years as ore-deposits of iron, copper, and zinc. The 

 phenomena point to a pneumatolytic origin of the ore-material, 

 and even the garnet may owe something to the same source,, 

 for, as Kemp has pointed out, it often approaches andradite 

 rather than grossularite in composition. 



I refer to such instances as these, chiefly in order to emphasize 

 their exceptional nature as depending upon a special kind of 

 composition of the original rocks affected. I cannot concur in 

 the large demands made by some geologists upon extensive 

 and radical metasomatism as, not merely a special incident, but 

 a general factor in metamorphism. This is not the place to 

 discuss so large a question, more especially since it turns in 

 particular cases upon the interpretation of field -evidence. I 

 would observe, however, that, when the appeal is to liquid solu- 

 tions emanating from igneous intrusions, speculation should be 

 controlled by regard to the known chemistry of crystallizing 

 magmas. The adinole transformation, for example, presents no 

 obvious difficulty on this score, since we know that the final 

 residual magma of an igneous rock may be very rich in albite. 

 The case is different when magnesian solutions of magmatic 

 origin are invoked to explain the dolomitization of a limestone, or 

 the production in it of such minerals as biotite and hornblende. 



What most of all characterizes the reciwstallization of rocks in 



