444 ME. A. WADE ON THE LLANDOVEEY AND [Aug. I9H, 



writer in the southern end of the May-Hill Inlier. 1 The Silurian 

 strata do, however, overlap the lower beds, each group in turn 

 creeping over the underlying group. 



The overlap of Powis-Castle Beds has already been traced across 

 the map. It is accompanied by the thinning-out of these beds 

 over the anticline of the Guilsfield Valley. This is first suspected 

 on account of the disappearance of the conglomerates and associated 

 beds between Welshpool and Y Frochas, is confirmed by the fact 

 that the conglomerate is only* represented by a very thin bed 

 (perhaps a foot or two thick) on the west side of the Welshpool 

 Dyke in the Ceunant gorge, and is absolutely proved by the outlier 

 of Wenlock Shales at Tyn-y-llywn, which rests directly upon the 

 Gaerfawr Grits with no intervening Llandovery Beds. Here, 

 therefore, the conditions are exactly the same as at Rhayader, 

 where Dr. Herbert Lap worth has accounted for peculiarities in the 

 distribution of the Llandovery (Caban) Conglomerate in the same 

 manner. 2 



It would seem, then, that the whole of the Welsh Border was 

 dotted with a chain of islands in Llandovery times, due to the fact 

 that folding along north-easterly and south-westerly axes had 

 already commenced. The Llandovery conglomerates were laid 

 down in the synclinal depressions, and this accounts for the rapid 

 changes noted in both the thickness and the lithological characters 

 of the beds over limited areas. The islands were submerged locally 

 in Wenlock times, and the overlap of the Ludlow, shown at the 

 southern end of the map (PI. XXXIII), which is probably true 

 overlap, indicates further subsidence at a later period. 



VII. CoEEELATION OF THE StEATA. 



In attempting to correlate the sequence with that of other areas, 

 I paid visits to other Welsh and Border districts — so as to 

 make my comparisons, as far as possible, dependent on personal 

 observation. 



(1) Shelve. — It would naturally be expected that a close simi- 

 larity would exist between the Ordovician succession in the Welsh- 

 pool area and that of Shelve, only 6 miles away to the south-east. 

 Nothing, however, could be more different. The Borrington Flags 

 may correspond to the Trilobite-Dingle Shales ; but the alternating 

 shales and ashes above them have no counterpart in the Welshpool 

 area, unless it be in the ashy grits which occur along with the 

 limestone development in the Gaerfawr Grits of Gwern-y-Brain. 

 These may correspond to the Whittery Group. 



(2) Car ado c. — Travelling 20 or 30 miles to the south-east, we 

 come to an outcrop of Ordovician rocks which agrees tolerably well 

 with the succession here described. The sequence in the Caradoc 



1 During field-work under Prof. W. W. Watts in 1907. 



2 ' The Silurian Sequence of Ehayader ' Q. J. G. S. vol. lvi (1900) p. 117. 



