182 MR. H. H. THOMAS OK THE [May I9II, 



Unfortunately, this is the only direct evidence as to the age of 

 the series that it is possible to bring forward ; but much important 

 information which bears indirectly upon this subject can be gathered 

 from other districts in the west of Pembrokeshire. 



At Musclewick, a little over a mile to the east of the map, near 

 the village of Marloes, the lower limit of the Skomer Series is 

 reached. Concerning this, Mr. T. C. Cantrill has kindly furnished 

 the following account : — 



' The lower limit of the Skomer Series is revealed between St. Ishmaels and 

 Marloes. A rhyolite, which here forms the lowest exposed member of the 

 Volcanic Series, is everywhere bordered on its northern or lower margin by an 

 outcrop of Llandeilo Flags from which it dips away without obvious discord- 

 ance. At first sight, therefore, it would appear as if the igneous series was 

 newer than, and immediately succeeded, the Llandeilo Flags, and might there- 

 fore be referable either to that formation or to some part of the Bala or Lower 

 Llandovery Series. Further examination, however, of the section exposed in 

 Musclewick Bay proves that there the junction of the volcanic rocks with the 

 Llandeilo is a faulted one ; but, unfortunately, the fault is vertical, and the 

 direction of movement remains uncertain. This section, therefore, still renders 

 it possible for the igneous rocks to be newer than the Llandeilo Flags, against 

 which they are faulted ; but general considerations make it probable that the 

 whole length of junction is a fault, along which the volcanic rocks as older beds 

 have been thrust from the south on to the newer Llandeilo Flags.' 



It appears, therefore, that while pre-Upper Llandovery in age 

 they may be even pre-Llandeilo. This suggestion is borne out by a 

 consideration of the sequence of Lower Palseozoic rocks exhibited 

 in the neighbourhood of Haverfordwest, about 10 miles away. In 

 that district there is practically an unbroken succession from the 

 Upper Llandovery to well down in the Arenig Series, and in this 

 sequence there is no indication of any marked period of vulcanieity, 

 neither is there any stratigraphical break of sufficient magnitude to 

 mark the eruption of some 3000 feet of volcanic rock in an area so 

 close at hand and along the same line of strike. Of course, we are 

 aware that both the beginning and the close of the Didymograptus- 

 murchisoni Zone and the upper limit of the Llandeilo Series were 

 horizons of volcanic activity in other districts ; but in South Wales, 

 south of the pre-Cambrian ridges of St. David's and Hayscastle, these 

 horizons are marked only by thin ashes, which show no tendency to 

 greater development in a westerly direction. 



In the Didymograptus-extensus Zone, however, we meet with 

 evidence of prolonged volcanic activity, as indicated by the great 

 thickness of acid and intermediate lavas at Trefgarn, 1 north of 

 Haverfordwest. The Trefgarn Volcanic Series, in the composition 

 of its members, presents many characters in common with those of 

 the Skomer Series, and the rocks evidently belong to the same 

 petrological province, although nob showing the variety of the 

 Skomer Series. 



1 The age of these rocks has lately been established by the discovery of an 

 Arenig graptolite-f auna in shales associated with the volcanic rocks in Trefgarn 

 gorge. See * Greology of the Country around Haverfordwest ' Mem. Ceol. Surv. 

 (in the press). 



