THE TILL. 541 



surface of the sandstone is rounded and retains everywhere the glacial 

 scratches perfectly. These are broad, deep grooves, uniformly directed 

 S. 30-40" E. On the south side the sandstone was uneven and ended 

 abru])tl\' in a nearly vertical wall, against whi(;h rested a mass of dark-gra\- 

 till of stony compactness, the surface of which was an exact continuation of 

 the broad, convex, striated surface of the sandstone, showing that the ice had 

 ])assed over them both together and planed them down to a common level. 

 Again, in changing the grade of the C!an;d Railroad, near the South 

 street bridge in Northampton, a section was exposed where tlie compact 

 stony clay abutted on the east against the I'ed sandstone, continuing the 

 curvature of the convex roche moutonnee surface of the sandstone in the 

 same way. In each case the drift and the sandstone were covered by 

 the Champlain clays in such a way as to show that the exact surface of the 

 drift upon which the ice rested had been covereil with the clavs immediately 

 after the disappearance of the latter, a point 1 have developed more fullv in 

 discussing the clays and their relation to the valley drift at the beginning of 

 Chapter XX. 



THE COARSE VALLEY DRIFT. 



On the west side of the river in Northampton the bay formed b}- the 

 retreat westward of the crystalline rocks is much deeper, the drift accu- 

 mulated there is more abundant, and the drumlins are on a larger scale. 

 The deposit is, however, much more masked by the later accumulation of 

 sand in the flood period, by which the whole surface is brought up to the 

 level of the highest terrace. While the deposit is probably the exact 

 equivalent of the valley drift alread}' described from the east side of the 

 river, I have thought it best to describe it separately, both because it 

 occurs in a separate portion of the basin and because it presents several 

 points of difference when compared with that. 



While the paste is clayey and well compacted, it is generally much 

 coarser, bowlders above a foot in length often making up three-fourths of 

 the mass, and masses above 3 or 4 feet in length being in places very 

 abundant. The three outcrops already described in the preceding section 

 from the west side of the river agree in their fineness and exceeding com- 

 pactness with the fine valley drift with which they are associated, and, like 

 it, certainly rest directly upon the older rocks. I have not been able to 

 ascertain if this was the case with reg-ard to the coarse valley drift here 



