Cole — The Iiifrimre Gneiss of TiverriU mid Druwahair. 365 



the crack which determined its position. The case is conclusive 

 against the production of the foliation in the gneiss hy subsequent 

 pressure. The banding is due to primary flow ; the metamorphic 

 effect of the hot magma on the amphibolite can be clearly traced ; and 

 the accumulation of garnets locally in the gneiss, at the expense of the 

 amphibolite, is easily observable in the field. 



The absence of marked alteration in the colour of the aplitic gneiss 

 indicates, however, that little absorption of basic amphibolite has here 

 gone on. I have elsewhere^ given reasons for regarding streaks and 

 layers of garnet in this type of gneiss in north-west Ireland as 

 distinctly derivative ; but we may conceive that the foreign material 

 absorbed in Castleore was a series of quartzites, schists, and lime- 

 stones, containing only a few basic igneous rocks. In this and similar 

 cases, we may picture the garnets as arising during the early stages 



Fig. 2. 



Dykes and trough-like intrusions of gneiss (fluidal aplite) in 

 amphibolite, Castleore. From a photograph by the author. 

 "Width of the block shown, 80 cm. 



of metamorphism of the invaded masses, and then being carried ofi, 

 and frequently dissolved, in the dominant intrusive rock, which in the 

 first instance promoted their growth along the contact-zone. In 

 other cases they may have formed a constituent of an already 

 metamorphosed and schistose series, into which a granitic magma pene- 

 trated, inducing the formation of sillimanite, altering amphibole to 

 biotite, but without effecting much else in the way of crystallisation. 

 In confirmation of the above observations, it should be stated 

 that in a section in the Geological Survey collection, cut from a granulitic 

 gneiss near Slishwood, the garnets are associated with patches of biotite, 



1 Op. cit,, Trans. E. I, Acad., vol. xxxi., p. 457 ; also p. 4.'i6. 



