ANCIENT WATER LEVELS OP CHAMPLAIN-HUDSON VALLEYS 113 



the terrace building by gravel-bearing streams. This evidence 

 is stronger when it is noted that both rock terraces and glacial 

 terraces are at this point in the river somewhat higher than at 

 Peekskill. The elevation of the glacial terraces at Peekskill is 

 from 100 to 120 feet; in the vicinity of West Point it is from 

 160 to 180 feet. Unless there has been a differential postglacial 

 uplift of the axis of the Highlands, this difference of level of 

 terraces at points about 9 miles apart, appears too great to be 

 explained by the normal tilting of the continent on the supposi- 

 tion that the deposits were originally made at the same water 

 level. If made, however, in ice-confined waters, their difference 

 of level is expectable. 



In the view of the terraces at West Point and Cold Spring 

 having been laid down marginal to ice filling the channel in the 



IFig-. 11 Cross-section of the Hudson valley at the West Point stage. A, West Point ter- 

 race near cemetery ; B, Constitution island; C, terrace south of Cold Spring-; D, Crow's 

 Nest mountain ; P,' ice at stage of the 400 foot morauie ; G, ice at West Point terrace stage 



manner of glaciers in the fiords of Norway, the lack of drift on 

 Constitution island above referred to is at once explained, 

 since it must have been at the time covered with ice, the cross- 

 section of the gorge then being that shown in the annexed figure. 



At Cold Spring on the south, facing Foundry cove, is a narrow 

 terrace, rising about 40 feet above sea level. 



Partial summary of preceding chapters. The front of the ice 

 sheet retreating northward from the terminal moraine and up 

 the Hudson valley halted temporarily at Tappan. The extension 

 of the ice east and west of this locality is as yet imperfectly 

 known. It certainly must have formed a broad sheet, rising on 

 the north, over Little and High Tor, and filling the canyon of the 

 Hudson in the Highlands if it did not also cover these last named 

 elevations. Northward, the broad valley of the Hudson was still 

 wrapped in the glacial sheet. 



At Haverstraw and Croton, evidence exists of a temporary halt 

 of the ice front, at a time when it had a rather marked convex 



