484 ME. F. W. HAKMEE ON" THE [NOV. 1907, 



suggest, may originally have flowed southwards towards Frome, 

 and thence into the great plain of Glastonbury through the gap 

 between the Mendip Hills on the one side and the Wiltshire Downs 

 on the other. 



No other case exists, over a region extending 150 miles from 

 south-west to north-east, in which water runs from one side of the 

 Jurassic escarpment to the other, or in which a channel through it 

 has been cut down to the level of the plains. 1 



The Avon leaves the Trowbridge basin abruptly, at a sharp angle 

 to the longest diameter of the latter, and to what may have been 

 the original direction of the drainage, entering near Bradford- 

 on-Avon a recent-looking gorge which connects the Trowbridge 

 plain with the Bath valley, one of a group of similar character, 

 but of a different and apparently an older type, which inosculate 

 near Bathampton. Like that at Ironbridge, the Bradford gorge 

 seems to form an artificial connexion between two separate basins. 

 Contour-maps give us no reason to think that it is a short iso- 

 lated remnant of some pre-Glacial drainage-system older than the 

 present physiographical features of the district, as, for example, 

 than the erosion of the Trowbridge plain and of the opening to the 

 south of it between the Mendip Hills and the Wiltshire Downs- 

 Its modern appearance is opposed to such a view. It cannot 

 have originated as a dip-slope valley : not only does the water run 

 through it the wrong way, but the general trend of the drainage 

 between Bradford and Bristol is in an opposite direction to that of 

 the dip. 



It is difficult to understand that the drainage of the Trowbridge 

 plain could have originally taken place across the Bradford ridge. 

 Valley-erosion would have more readily followed the strike of the 

 Oxford Clay to the east of Frome. A possible explanation of the 

 facts seems to be that the Bradford gorge originated as an over- 

 flow-channel to a Glacial Lake Trowbridge. The excavation of the 

 former, on this view, however, could not have commenced while the 

 col south-west of Frome stood at its present level; but drainage 

 through the latter might have been obstructed by sedentary ice 

 accumulated in the col, or in the Bristol Channel, as before 

 suggested. In the latter case, a still larger lake must have existed 

 over the Glastonbury depression. Bartholomew's maps show, in an 

 interesting way, possible overflow-channels from such a lake: by 

 Wellington into the valley of the Exe, by Ilminster into that of 

 the Axe, or by Yeovil into the Stour. 2 



If the view here taken be correct, the raising of the water-level 

 of the Trowbridge plain, sufficiently to have caused the excavation 

 of the Bradford gorge, would have carried the suggested lake over 

 the Malmesbury-Swindon ridge into the basin of the Thames ; but 



1 The gaps in the Jurassic escarpment at Lincoln and Ancaster will be dis- 

 cussed later on (§ VI, p. 486). 



2 These channels lie south of the region represented in PL XXXII. 



