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



abundant of these substances, though others, such as boric and 

 hydrofluoric acids, are more potent. Others again, such as 

 carbon dioxide, may acquire importance in particular cases. Now 

 these are precisely the volatile bodies which we know to be 

 present in plutonic magmas, and to be given out after fulfilling 

 their functions in the crystallization of plutonic rocks. Various 

 petrogenetic considerations show that the quantity of these gases 

 in an intruded magma is very considerable, and almost the whole 

 must pass into and through the rocks invaded. This is true of 

 all kinds of magmas, even when the resulting igneous rocks, such 

 as peridotites, preserve no direct evidence of the fact. There can 

 be, I think, no doubt that the solvent medium wtiich is essential 

 to metamorphism is to be ascribed ultimately to a magmatic 

 source. Since, however, metamorphism can evidently go on far 

 from any igneous intrusion, we are to conceive a pervading 

 medium, doubtless of extreme tenuity in general but attaining a 

 greater concentration in the neighbourhood of newly - intruded 

 igneous rocks. Numerous facts suggest that other volatile bodies 

 besides water do permeate the earth's crust at large, if only in 

 an exceedingly dilute state. For instance, Carnot's observations 

 go to show that fossil bones undergo in the course of geological 

 ages a slow enrichment in fluorine, until the phosphate attains 

 to the composition of apatite. 



The diffusion of the volatile solvent here premised can become 

 effective, despite the extreme slowness of the process, because it 

 is perennially in progress. Diffusion of dissolved material, on 

 the other hand, can take place only during the time of meta- 

 morphism : that is, while the requisite conditions of temperature, 

 etc. are maintained; and it is in consequence restricted within very 

 narrow limits. We find that in a banded sediment composed of 

 different thin seams — argillaceous, calcareous, and gritty — meta- 

 morphism does not confuse the whole to one average t}^pe. The 

 several bands remain distinct, each represented by its own asso- 

 ciation of new minerals, even when a dozen such bands are 

 included in the field of the microscope. Apart from the effects 

 of mechanical disturbance, phenomena indicative of segregation, 

 such as garnet-lenticles and streaks of quartz sillimanitise, 

 are found only in the highest temperature- grade of metamorphism. 

 Here, too, a coarser grain of the recrystallized rocks points to a 

 somewhat enlarged amplitude of diffusion, the diffusion-constant 

 being of course a function of temperature. The free circulation 



