Prof. Blake on the Volcanic Group of St. Davids. 323 



the appearances are due to the intrusion along a fault of a diabase 

 dyke which has caught up large fragments of granite, felsite, and 

 conglomerate, and cemented them in its substance ; but the granite 

 scarcely anywhere comes in contact with the conglomerate, and is 

 nowhere intrusive. The junctions in the Allan valley are all faulted 

 with forked faults, some reversed, others normal, the intervening 

 mass often decaying. 



The supposed isocline west of the granitic mass cannot be verified 

 on an examination of the coast-section, there being great irregularity 

 and gentle synclinals not far from where the apex of the isocline 

 should be. 



With regard to the nature of the rocks which thus antedate the 

 Cambrian, the author was unable to recognize any true alternations 

 in the materials of the granitic axis, though the rock is a peculiar 

 one in the arrangement of its constituents. The felsitic rocks are 

 not independent of the granite, as they surround it on all sides, the 

 line along the north and south being specially traced. They are 

 also often intrusive into the ashes, and hence can have no definite 

 strike. The general features of these rocks are therefore most 

 easily to be matched in such volcanic districts as that of Mull. 



These results are confirmed by the structure of two outlying 

 masses, one at .Ramsey Island, the other south of Points Castle. In 

 the former, at Forth Hayog, quartz-porphyry is succeeded by a band 

 of rhyolite showing flow-structure, and this by a^hes and agglo- 

 merates. On the South Cam is a mass showing perlitic structure 

 and contorted lines of flow, and on the north, at Pwll Heudre, large 

 masses of banded spherulite of somewhat doubtful character, and an 

 apparent strike of E. and W. The conglomerate in this area has 

 many pebbles of the associated rocks. In the latter area we have a 

 similar series separated by a fault from the Cambrian rocks, and 

 consisting of banded felsites, weathering into apparent beds of very 

 irregular lie, and followed by an ashy series. 



The mass which lies to the east and terminates at Roche Castle 

 does not present sufficient similarity to these to be included in the 

 same description. The Koche-Castle rock, however, instead of being 

 bedded, was originally andesitic or trachytic, the felspar crystals 

 having been replaced by pseudomorphous quartz. 



Attention was drawn to the highly acid character of the whole 

 series, and to the small size of the centres of eruption, and it was 

 suggested that such centres have continually decreased in number 

 and increased in magnitude during geological time. 



9. " On further Discoveries of Vegetable Eemains in the Triassic 

 Strata of the South Coast of Devonshire, between Budleigh Salterton 

 and Sidmouth." By A. T. Metcalfe, Esq., E.G.S. 



