Vol. 63. j THE ORIGIN OF CERTAIN CANON-LIKE VALLEYS. 483 



The present case is interesting, however, as it closely corresponds 

 with that of the Severn near Bridgnorth. It will be instructive, 

 moreover, to compare the abrupt way in which the Camlad leaves 

 the lower ground of the Wenlock Shales to cut into the higher 

 Ordovician rocks, with the behaviour of the Avon when entering 

 the Clifton Gorge near Bristol, to be dealt with in the next section. 



The mutual resemblance of the phenomena described in this 

 paper, even in small details, is remarkable, and the argument as to 

 the mode of their origin will be found to have, I think, a cumulative 

 force as we proceed. 



It will be noticed that the level of the site of the supposed lake 

 is lower than that of the ground north of the junction of the two 

 streams ; the lake seems to me to have been deepened pari passu 

 with the excavation of its outlet. This feature, common to all the 

 regions here discussed, appears to support the view that the erosion 

 of such basins was due to lacustrine rather than to fluviatile 

 action. 



Y. A SUGGESTED LAKE TROWBRIDGE, AND THE GORGES AT CLIFTON 

 AND BRADFORD-ON-AVON. 



There is another large lake-like depression in the West of 

 England, in Wiltshire, the special features of which also correspond 

 closely with those of the Pickering region. 



PL XXXII, a contour-map of the surrounding district, which 

 should be compared with that of Lake Pickering, shows the basin 

 in question to be of irregular shape, about 30 miles long and from 

 5 to 12 miles broad, very nearly of the same size as its Yorkshire 

 prototype. 



Bounded on the east by the high Chalk-land of the White Horse 

 Hills l and the Wiltshire Downs, and on the west by the Jurassic 

 deposits of the Cotteswolds, it forms part of the great Oxfordian 

 plain which extends diagonally across England from Wiltshire to 

 Buckingham and Winslow, and thence, somewhat higher land 

 intervening, into the Eenland. 



The western part of this plain is divided by a low ridge between 

 Malmesbury and Swindon, separating the catchment-areas of the 

 Thames (or Isis) and the Bradford Avon. 2 For reasons given here- 

 after, I think that this ridge may have formed the water-parting 

 in pre-Glacial times ; there is no indication that any important 

 valley formerly joined the two areas, or that any part of the Oxford 

 plain drained westwards into the Bristol Channel. 



North-east and south of the ridge, basin-shaped depressions exist 

 without natural outlet, associated with canon-like channels of a 

 modern type excavated laterally across higher land — the Thames 

 basin being drained by the Goring gorge, and the Trowbridge basin 

 by that of Bradford-on-Avon. The drainage of the latter region, I 



1 This is the name adopted for the Berkshire Downs by Dr. H. E. Mill, 

 Geogr. Journ. vol. xxiv (1904) p. 634. 



2 Not to be confounded with the Warwickshire or the Wiltshire Avon. 



