Glaciers and Glacid Radiants — Chiypolt'. 77 
The direction above mentioned, uainely fiom the norfclj-east, 
is that in which the general flow of the ice over the midland 
district might have been expected to occur, if we allow due 
weight to the datum stated above. Granting a heavy precipita- 
tion in the north-east of the continent, where the land was also 
high, it is natural to expect that tlie ice wonld move off toward 
the valley of the Mississip]>i where the land is low. Accord- 
ingly we find the greatest southward extension in the states of 
Indiana, Illinois and Missouri, Farther west the line of the 
extreme south limit of the ice rapidly recedes to the north until 
it nears the boundary line or perhaps altogether retires into 
Canada. 
This is in perfect harmony with the fact that this interior 
northern region is now the region of least precipitation. The 
same was most likely true at the time in question. The vast 
mass of the ice would be formed in the north-east and its quantity 
diminished to the we.st and south. 
We may therefoi'e regard the higli lands of Labrador and the 
vicinity of Hudson's bay as the great gathering ground of the 
eastern ice-sheet which flowed away to the south-east, south and 
south-west in the way above described. 
Data are yet scanty regarding the region to the north of this 
district. Extreme writers have taken it as a matter of course 
that it was covered with a sheet of ice creeping down from the 
area round the pole. But this has been for the most part a 
matter of inference or of assumption. Granting that in all 
probability the Atlantic border was the region of great precipi- 
tation here as in other parts of the continent it is nearly certain 
that the snowfall diminished inland, so that this region was 
under conditions similar to those that prevailed farther south. 
Greenland was doubtless then as now a glacial radiant. The 
wild and desolate strip of land between Lancaster sound and 
Hudson's strait was another great gathering ground. But the 
vast polar archipehigo around Melville sound was nearly 
in the condition of the midland plain of Canada and the states, 
and afforded less material for glacier-making. There seems to 
be consequently little basis for the construction of an ice-sheet 
of enormous thickness in this region. 
In confirmation of this assertion we find from the geological 
reports issued by the Canadian survey that the indications are 
