THE SEVERN AND ITS TRIBUTARIES. 623 



find here a Scrobicularia-cla.j, occurring at the same depth as in 

 the shaft at Caldicot. near Portskewet, which is about nineteen 

 miles further north. The clay rests on gravel consisting of Millstone- 

 grit pebbles. 



No. 10 is taken from Mr. Evan D. Jones's paper (loc. cit.) ; it was 

 obtained in sinking a shaft through Caldicot Marsh near Port- 

 skewet. In Mr. Jones's paper the section has slipped too near the 

 datum -line in the figure ; it should be raised till the surface of the 

 clay is on the same level as that of no. 6. fig. 4. The Scrobicularia- 

 marl which occurred in this section is of yellowish-grey colour, and 

 is characterized by an admixture of freshwater and brackish-water 

 shells such as Limncea, Planorbis, Scrobicularia piperata, and Car- 

 dium edule. Diatoms are common in it, and also remains of 

 Chara. It presents a striking contrast in colour and composition to 

 the rest of the alluvium. 



The foregoing material enables us to classify the alluvial deposits 

 in the following manner : — 



f a. More sandy zone, 5 to 7 feet. 

 Zone 1. Upper clay. - b. More argillaceous zone, with disseminated vegetable 

 ' [ matter. 



Upper peat. 

 Zone 2. Lower clay 

 Lower peat. 

 Zone 3. Sands and mud. 

 Gravel. 



With the gravel we have not now to do ; the presence in it of 

 glaciated stones is very suggestive ; but, as further information 

 with regard to it will be forthcoming as the works of the Severn 

 Tunnel progress, we may defer for the present the question of its 

 formation. The sand is evidently a marine or tidal deposit, like 

 the blue clay, from which it differs chiefly in containing no 

 argillaceous sediment or scarcely any. It is not a constant 

 deposit, and may possibly be explained by local current-action, 

 though the gradual succession from coarse through fine gravel to 

 this sand in ascending order seems to require some more general 

 explanation. The fact that it is entirely absent at Caldicot on the 

 opposite side of the river, but is represented there by ordinary 

 blue silt, is in favour of merely local action. The mud below this 

 gravel in Section 3 is on this horizon ; its upper part, on which the 

 gravel reposes, is penetrated by rootlets, indicating a land surface. 



There is a difficulty, when only one layer of peat is present in a 

 section, in determining whether it is upper or lower ; we may safely 

 regard the lower bed at Porlock, and at New Passage, and the single 

 bed at Lydney as lower peat ; the remaining peat-beds are probably 

 upper. 



The lower peat must evidently have accumulated at a time when 

 the relative level of land and sea was different from that of the present 

 day. At New Passage, for instance, it lies at half-tide level, and would 

 have been, under existing conditions, always submerged for one 

 half the day. An elevation of twenty feet, however, is all that is 



