612 GEOLOGY OF OLD HAMPSHIEE COUNTY, MASS. 



of water which flows out from under one of the Greenland glaciers; then, 

 like the great river Yukon, with its vast volume and lakelike expansions, 

 and finally, like the Rhone, heading in a glacier among our smaller New 

 England Alps. As Ave trace around the valley the great bench which 

 represents the completed work of the chain of lakes, we must imagine the 

 waters up to and over its level, the tributaries entering at its outer border,^ 

 and the deep notches they have cut in the bench, as in later time they 

 followed the lessening waters toward the center of the basin, again filled to 

 its level. We must restore also many smaller gulches cut in the bench 

 where there are now no running streams. We must be on our guard for 

 places where, from want of material or the rapid flow of the waters in con- 

 tracted places, the bench was not built up to the full level, and, most 

 difficult, must try to form some estimate of the amount carried away or 

 rearranged by the river itself as it sank from its greatest height to its 

 present level. We must inquire also whether kame material — stratified 

 beds formed while the ice still held the waters above their normal level — 

 does not blend with and disguise the true terrace of later time, and from 

 a study of the inner structure of the beds must seek to learn whether it 

 rose gradually to, or asserted from the first, its highest level. 



In no part does the map need more to be supplemented by sections 

 than here, for it can show only those portions of the lake-shore deposit 

 which have escaped degradation and of the lake-bottom beds which have not 

 been molded anew by the river into terraces of a later time and covered by 

 the layers of river-bottom sands and loess (t--t"', PI. XXXV), which have 

 for the most part thinly concealed rather than replaced the more massive 

 deposits^ of this age. The word "terrace" — which we employ in geological 

 discussions with a latitude of meaning for the most part useful but some- 

 times liable to be misleading or indefinite, and which from the great develop- 

 ment of this structural form along the Connecticut always suggests here the 

 ordinary river-erosion terraces — I am inclined to replace by the word 

 "bench " in the following descriptions, to emphasize the many points of 

 distinction between the highest level as compared with the remaining 

 levels we pass over in going down from the high ground to the river. 



If we take a single terrace lower in the series for examination, we shall 

 find it a plain sloping with the river and bordered on the side away from 



' Outside 1 8 h on PI. XXXV. 

 -SeeG. K. Gilbert, Outlet of Lake BoQDeville: Am. Jour. Sci., 3(1 series, Vol. XIX, 1880, p. 341. 



