PKE GLACIAL EROSION. 515 



River." It was temporarily reoccxipied dui-ino- the recession of the ice, 

 receiving- the ovei-flow of a Glacial lake wliich formed on the north flank of 

 the Holyoke range, banked on the north by the ice of the Hadle}' Ijasin. 



There are two striking gorges in the west of the town of Holyoke, 

 both cutting the traj^ very obliquel}-, one occupied by Wrights Brook 

 (which enters Hitchcocks Pond), while the Westfield and Holyoke Rail- 

 road passes through the other. These gorges seem to be portions of the 

 bed of a stream that gathered on the east flank of Mount Tom and ran 

 south into the Westfield River. 



Another notch of unknown depth cuts the trap ridge just where it 

 crosses the State line into Connecticut. This I have connected with the 

 large brook which comes down from Sodom Mountain, in Granville, and 

 have called it on the map the Southwick notch. 



Though the evidence is much less clear, it seems probable that the 

 narrow canyon skirting the east front of Mount Toby was cut by Locks 

 Brook. Its bottom has now the shape of an abandoned water channel. It 

 is probable that the j^ortion of the channel of Locks Brook which ran in 

 sandstone between the end of its gorge in the crystalline rocks and the 

 beginning of the canyon was removed b}- ice erosion. At the end of the 

 Glacial period the ice, halting in the Montague basin, deflected the brook 

 again southward into this canyon. 



THE CHARACTER AND AMOUNT OF EROSION DURING LATER MESOZOIC TIME 

 AS COMPARED WITH THAT OF THE GLACIAL PERIOD. 



From the preceding discussions of the crystalline rocks and the Tri- 

 assic sandstones it is certain that the broad Connecticut Valley was an 

 orographic feature of first importance formed in the crystalline rocks before 

 the deposition of the sandstones, its borders coinciding closely with the 

 present boundaries of the latter. Prof. W. M. Davis^ has suggested that 

 there may have been a pre-Triassic peneplain over this area. The places 

 where the crystalline rocks break through the Trias are at such diff'erent 

 levels in places very near one another that this is not probable. 



This valley was then deeply filled by the sands of the Trias, indeed 

 above and beyond the present lips of the basin, and has been since so thor- 

 oughly eroded a second time that only remnants of this filling remain. It 

 seems quite certain that the walls of the valley during and at the close 



' Bull. Geol. Soc. America, Vol. II, p, 549. .Jour. Geol., Vol. IV, p. 678. 



