40 Observations on the 



tic forms, especially cubic, and columnar. But in this 

 respect it is imitated by aged bodies of dun limestone, which 

 at the New Weir, form huge wall-like parietes to the hills, and 

 more particularly so at WindclifF, and in other situations, 

 like the lonely fragments of immense structures, or as druidical 

 ruins covered with lichens and infant oak-shrubs. It can hardly, 

 be supposed that the flow of springs has formed fissures and 

 burst these rocks. They appear to have been exposed, disunited 

 and precipitated by the operation of torrents washing away the 

 soils on their sides. Enough are left in parallelism to evince 

 their primitive position ; we find them increasing in abundance 

 and disunion, lying over the level surface of the hills, above the 

 valley of Troy-mitchell, towards Chepstow. On the common 

 of Trelech, from accidental circumstances of arrangement, they 

 lie over upright stones, like funereal monuments, or cromlechs, 

 and on this spot rather d propos, for the gieat contest of 

 Harold with the Welch ensued here. The pebble which mingles 

 principally with this aggregate rock is white quartz, which gives 

 lio-ht on attrition ; I have found one or two varieties with diffi- 

 culty to be discriminated from Carnelian, being very translucent : 

 very pure red quartz also occurs, and has doubtless been scat- 

 tered from the debris of these rocks about the general surface of 

 the lower soil and the bed of the river. 



We next come to the new sandstone. Of the quality of this 

 rock I know very little : it is a finer combination of sandy par- 

 ticles than the old, much lighter in colour, and more seldom 

 found. It has been tried in building with expectations which 

 are very far from being verified, probably from its being mingled 

 with calcareous matter. 



Above this we find more particularly the marls, red and 

 green. This name is uniformly given and very improperly to 

 these thin talcy stratifications. On analysis I find them to be a 

 ferruginous composition of sand and clay, with about eight 

 grains in an hundred of calcareous matter. Some have sup- 

 posed them to have been formerly sandstone, and an old rock 

 at Newnham on the banks of the Severn, almost entirely con- 

 sisting of it, has been conjectured, particularly to have under- 

 cQ}ie such transition ; but it is to be remarked that I have found 



