32 
KENDALL : THE GLACIER LAKES OF CLEVELAND. 
which flow westward and eastward from a scarcely perceptible 
watershed across the middle of the valley. These cannot explain 
the valley in which they flow, and it seems best accounted for 
by the diversion of the Derwent by the ice-sheet through the 
glacial overflow of Hackness Gorge, but that previously the 
Derwent flowed down the Symes Valley to the sea at Burniston 
(Plate XIV.). 
The explanation of this striking change in physical geography 
was the impounding of the drainage of Harwood Dale by an 
ice-dam across the Symes Valley which formed a lake, one arm 
of whicli stretched up the Jugger Howe Valley, and the other 
«p Lownortli Valley, which found an overflow across a col, and 
Fig. 9. 
SECTION ACROSS JUGGER HOWE BECK AND CASTLE BECK, 
SnOWIN(! THE MORAINE. 
out a steep gorge, Langdale, into the Hackness Lake. This 
gorge was cut so rapidly that before the retreat of the ice it 
was established as the line of permanent drainage, and any 
resumption of the old channel to Burniston was rendered im- 
possible. 
Lake Staintondale and Lake Hayburn. — The line of maximum 
ice-extension was not, however, long maintained. The margin 
•appears to have been withdrawn from the line of Jugger Howe 
Beck rather rapidly, until it fell behind the water-parting of 
Pye Rigg and Hollow Bigg, beyond which the retreat became 
much more gradual, with long and frequent halts and some re- 
advances. Marginal lakes then were formed, and two are 
^especially definite, one at the liead of Staintondale and the other 
