GOLDTHWAIT: SAND PLAINS OF GLACIAL LAKE SUDBURY. 265 
rangement of deltas used by Professor Crosby and others; and finally, 
applying this scheme to the data from the Sudbury valley, I shall show . 
that the sand plains of Lake Sudbury do not fall into flat steps, but 
probably into a tilted step-system. 
Conditions of Drainage along the Front of the Retreating 
Ice-Sheet. 
When the ice-sheet melted back over New England, uncovering a 
land surface of gentle irregularity, the country along its front must have 
met with temporary conditions of drainage quite unlike those of the 
present day. The ice must have blocked the courses of many streams, 
turning them one side or damming them up to form temporary lakes. 
The melting of the ice itself doubtless contributed in a measure to the 
volume of the streams in front of it, although the rate of melting may 
have been so slow that there was only a moderate flooding of the streams. 
Where the ice-sheet lay with its front along a divide between two 
stream systems, or on ground that sloped away from it, there was oppor- 
tunity in a given time for either an unusual amount of erosion or an 
unusual amount of deposition. Each stream that drained away from the 
ice would from the beginning attempt to make an evenly graded bed for 
itself, either degrading or aggrading according to the conditions of load, 
volume, and slope: the greater the supply of waste, the stronger the ten- 
dency to aggrade ; the steeper the slope, the greater the power to erode. 
An ice-fed stream which was given little waste to carry, and which ran 
down a steep slope, would be very active in eroding, while another stream 
which received an abundance of waste gravels, and which ran down a 
gentle slope, would be active in aggrading. In the first case, the result- 
ing physiographic feature might be a slope or a channel swept clean of 
all but boulders, coarse gravel, and bed-rock ledges ; in the last, it might 
be an alluvial fan or a waste-filled valley. 
As soon as the ice had melted back of a divide, so that its front lay 
across the lower part of a valley, it enclosed a basin whose margin was in 
part the ice-wall and in part the high ground which formed the water- 
shed of the valley. Inasmuch as each of these basins was originally 
covered by the ice-sheet, and was uncovered by its melting off, each must 
always have contained water up to as high a level as the lowest point 
on its rim; for each of these low points, or “cols,” would have served 
as an outlet to drain off the excess of water in the adjoining basin. 
Besides the water derived from the melting ice, whatever water was shed 
