TIME DIVISLONS OF THE ICE AGRE, 405° 
through a duration of probably one or two million years, 
which would permit the valley to have its origin chiefly by 
rock solution, as the deepening of the excavation to 200 or 
250 feet would be at an average rate of no more than an 
eighth or a third of an inch for each century. 
When the more rainy and snowy climate of the Glacial 
period caused larger floods of the streams, especially during 
the spring months, the residuary loam and gravel mantling 
all the surface were subjected to exceptional denudation. 
Considerable material was swept down every stream course 
and ravine, until, on debouching into the main valley, the 
flood could expand more widely and so lose its velocity and 
transporting capacity. The gravel and sand were then de- 
posited along the border of the broad bottomland and in 
alluvial fans upon the lower part of the inclosing slopes 
wherever tributary streams or the rills of rains and snow- 
melting descended. 
Instead of forming continuous, level-topped terraces, at 
successive heights, up to nearly equal vertical limits on each 
side of the valley, like the terraces of modified dmft on the 
Connecticut River along its upper half where it is the 
boundary between New Hampshire and Vermont, or like the 
usually less numerous terraces of this kind in most valleys of 
glaciated countries, the Scmme gravel and sand are of less 
amount, and have gentle or steep slopes toward the centre 
of the valley, presenting a terrace escarpment only where 
marginal parts of these deposits have been later carried 
away by the undermining action of the streams which 
brought them, or of the main river. Along the Connecticut 
River a wide flood-plain of modified drift was built up, filling 
the valley to the level of the highest continuous terraces, 100 
to 200 feet above the river; and the terraces are remnants 
of that flood-plain, and of the lower temporary levels 
occupied and abandoned by the river during its process of 
removal of the greater part of that original deposit of the 
modified drift. In the Somme valley, the supply of material 
was less than that set free from the melting North American 
ice-sheet, and it was insufficient to build up a flood-plain in 
any part of this valley below Amiens, though at many places 
it formed extensive deposits on either side, which sometimes 
reach, with slopes nearly like those inclosing the valley, 
from the bottomland up to heights of 50 to 100 feet; and 
patches of similar gravel and sand occasionally are observable 
also at greater heights, nearly to the verge of the uplands. 
