THE LAKE I5ENCH NORTH OF HOLYOKE KANOE. 649 



This broad, flat, sand- covered plateau in eontiiuiation of the soiitli eml 

 of Mount Warner has the exact height of the old Hadley Lake. Its direction 

 (southeast) was a great jiuzzle to me, and I tried to explain it l)v su])posing 

 that the south current and the west wind ])r(iduced a resultant southeast 

 direction in the great sand spit. Recently (18S,S) excavations along the 

 road south from the Catholic cemetery have shown that all along the south- 

 east front of the plateau the till lies almost at the surface and makes the 

 explanation more probable that the whole mass of the deposit is due to 

 ice, and that the north-south valley movement of the ice is here, where 

 the valley is unusually wide and open, replaced by the usual upland 

 (N. 30° E.) movement, and this agrees with the strong pressure of the ice 

 along the west face of Deerfield Mcjuntain. Only the surface and slopes of 

 the plateau were then molded later by the water and covered and flanked 

 bv sand bars. 



THE LAKE BENCH ALONG THE NORTH SLOPE OF THE MOUNT HOLYOKE AND 



MOUNT TOM RANGE. 



I have already (p. 586) called attention to the fact that great masses of 

 irregular sands are in places heaped up against the flanks of these ranges at 

 heights much above the highest water level of the Hadley Lake. Where, 

 as along south of Amherst, the high teiTace is a bench cut in these sands it 

 sinks gradually, and often without any marked change of slope, into the 

 lake bottom, as if there had been here no marked current, but an undertow 

 had drawn the sands in large quantity down into the deeper water. 



Farther west, south of Hadley and in the Holj'oke notch, the current 

 was more marked; but the matei'ial at the disposal of the stream was less 

 in amount and the terrace is a narrow bench, often of till, and from the 

 entrance of the notch dowii to Titans Pier the waters cut back the till in a 

 broad bench and then wore into the trap and sandstone, producing a \er- 

 tical wall which the talus of fallen trap has not yet obliterated. 



Across the river the same conditions hold. Above the highest terrace 

 level, as determined by its coincidence with the Florence plain, higher 

 levels of coarse sand occur and the lake bench slopes inward to where it is 

 cut ofi' by the later erosion of the Connecticut, or when we get beyond this, 

 as in Easthampton, it continues its gradual slope to the middle of the basin, 

 or to tlie line of tlie deenest water of the broad stream which flowed down 



