part 1] ANNIYERSAKY ADDRESS OF THE PRESIDE>'T. IxXV 



intrusive rocks are only in part of strongly sodic nature. The 

 intrusions are of later date than the volcanic outj)Ourings — some in 

 the Lough Mask area are post-Llandover}'- — and a change in the 

 sti'ess conditions ma}^ be presumed. We shall see later that calcic 

 intrusive rocks may be found in closer relation with volcanic rocks 

 of alkaline nature, and that the seeming anomaly is in accordance 

 with the principles laid down. 



In the south the submarginal belt of activity includes the 

 voleaiiic districts of Pembrokeshire, our knowledge of which has 

 been much enlarged in recent years. The lavas of Skomer Island, 

 described by Dr. H. H. Thomas, include a range from soda-rhyolites 

 through various types of soda- trachytes to mugearites and olivine - 

 basalts. To the same series we may probably refer the soda- 

 rhyolites of Trefgarn, near Haverfordwest, and perhaps the volcanic 

 group of Llangynog, near Caermarthen. This is described as 

 consisting of augite-andesites and rhj^olites ; but the former seem 

 to be of mugearitic rather than andesitic affinities, and so make no 

 exception to the general alkaline facies. 



These various occurrences are probably all low down in the 

 Arenig ; but those of North Pembrokeshire belong to higher hori- 

 zons, indicating a progressive displacement of the theatre of 

 activity towards the interior of the geosj^ncline. The ceratophyres 

 of Abercastle are intercalated in the Tetragraptus Beds, while the 

 soda-rhyolites and rhyolite-tuffs of Llanrian are in the upper part 

 of the Didymograptus-hifidtis Zone, and the fuller development at 

 Fishguard is probably at about the same horizon. Here, in 

 addition to soda-rhyolites, Dr. Cox has recognized typical spilites. 

 They were succeeded by basic intrusions, mostly in the form of 

 irregular sills. In these the stronglj^ alkaline nature is sometimes 

 less pronounced, but the dolerites of Fishguard have an oligoclase- 

 andesine as their dominant felspar, and the thick sills of St. David's 

 Head, probably of the same series, show albitization. 



In the interior of the British geosyncline late Cambrian and 

 early Ordovician eruj^tions were not wholly wanting, bu.t they were 

 on a feeble scale, and were confined to the neighbourhood of areas 

 which had already acquired special tectonic importance. If any 

 outbreak occurred on the borders of the Archaean horst of 

 Anglesey, the evidence of it has been lost. But the Harlech 

 anticline had alread}^ declared itself, perhaps as the reassertion of a 

 concealed Archaean centre, and on its skirts the first effects of 

 vulcanism were localized. The lavas in the Upper LingiiJa Flags 



