PRE-GLAOIAL EROSION. 517 



1 think it may be deduced from the tacts given above that the greater 

 portion of this erosion was performed by the first agent, but that the ice 

 wore into the soft sandstone considerably, and in some places enormously; 

 so that, if the Pleistocene deposits were removed from the valley, the rocky 

 floor bek)w would bear small resemblance to the surface upon which the ice 

 began to act. I imagine that the present surface of these latter deposits 

 would much more nearly coincide therewith. Indeed, along all tlie west 

 side of the valley from Deerfield to Southwick and beyond, and nortli 

 of the Holyoke range, the sandstones may well have been considerabU- 

 higher than the present cultivated surface of the valley. This is deduced 

 from the consideration that if the present drainage represents closely the 

 pre-Glacial, as shown above, the sandstone should rise by easy slopes from 

 the streams and be highest in the areas between them, or in some wa}' show 

 an intelligible relation to them. But from this point of view the deej) 

 depression in the sandstone west of the trap ridge in Deei-field and north 

 and west of the Holyoke range would render such a drainage impossible, 

 and must be a later work, which can only have been done by the ice. This 

 exceptional erosion of the ice depended largely upon the soft natiu-e of the 

 sandstones and the peculiar position of the trap ridges. 



From the top of Mount Holyoke I have seen the valley fog rest 

 against the hills east and west and, rising to my feet, spread, with a surface 

 level as the sea, up and down the valley as far as the eye could reach. If 

 it had risen a few huncb'ed feet higher I believe its mass would have 

 rudely equaled the pre-Glacial erosion of the Triassic, while I imagine the 

 present Pleistocene deposits in the valle}' would scarcely equal the amovuit 

 removed by the ice. 



As for the crystalline rocks which flank the broad Connecticut Valley 

 on either side, the fact that the newer crystallines are covered by the Trias 

 in the bottom of the valley and yet are abundantly present in the coarse 

 Triassic conglomerates, while the older Cambrian gneisses are broadly 

 exposed on the east but are not represented in the adjoining Triassic con- 

 glomerates, shows that there has been large erosion over tlie eastern plateau 

 since the Trias. The suggestion of Professor Pumpelly that secular disin- 

 tegration mav have deeply prepared these rocks for glacial erosion must 

 be taken account of, and renders it impossible to assign to pre-Glacial and 

 Glacial agencies their proper share of work. 



