432 CARTER : GLACIATION OF DON AND DEARNE VALLEYS. 
lake over Magnesian Limestone and through the Balby valley, 
which would sweep clean the sandstone bed of the valley (Plate 
LIX., Fig. 2) before it was invaded by the second or Stainmoor 
glacier, which would catch up and carry forward much of the Coal 
Measure and Magnesian Limestone material left by the lobe of the 
Pennine glacier. Another forward movement would bury the 
Balby valley, and the drainage from the lake would have to travel 
along tlie edge of the ice, and thus the Loversall valley was 
commenced (Fig. 9). This wa.s cut sufficiently low to take the 
drainage from the Edlington area when the ice retreated, and so 
the stream which starts towards the Don makes a semi-circular 
bend, and, doubling back parallel to its previous course, flows into 
the Trent (Fig. 3). Continued advance of the second glacier 
to Edlington would almost close the outlet of the great lake, 
but there is a shallow dip at Edlington very plainty shown on 
the six-inch map, as it just cuts through the 300-foot contour 
and is at right angles to the neighbouring valleys, which may 
be the slipJlow channel cut along the ice-front just before the 
Ipoke was closed (Fig. 9). When the ice-front reached Clifton 
(PI. LXL), the gorge would be entirely closed, and a lake thirty 
miles long would be formed, reaching from Clay Cross and 
Hardwick Hall on the south to Bretton Pa,rk on the north, and 
ramifying far into the tributary valleys of the Don and Pother. 
This lake would overflow by the Kiveton gorge into the River 
Ryton, and thence into the plain at Worksop. 
The gr2,dual advance of the ice up the southern slope of 
the Conisbrough gorge tow^ards Clifton, with the continual flow 
of water along its front, would explain the remarkable recession 
of the Upper ^[a,gnesian Limestone on this side, as compared 
with the opposite side of the gorge. It also accounts for the 
curious triplet of va,]leys at Hemsworth, Balby, and Loversall 
(Figs. 3 and 9). The outlet at Kiveton and the consequent 
level of the great lake at 330 feet (PI. LXL) also coincides with 
the w^arp bands between the boulder-clays at Staincross, wiiich 
lie just above the 300-foot contour (PI. LIV.). 
The second (Stainmoor) glacier is assumed to have advanced 
in this area until its front occupied a curve (PI. LXI.) from 
Staincross to Ardsley, and thence to Adwick and Conisbrough, 
