THE WARE A^TD SWIFT IIIVER LAKES. 571 



obliquely across the high ridge that separates the narrow "West Brancli 

 Valley on the west from the broad tiat-bottoiued Urauge-Enfield basin on 

 the east. 



The effect of the ice front as it rested here is seen clearly in the 

 filling- of the Beaver Brook channel (in the northeast corner of the Palmer 

 quadrangle and crossing the middle of Ware), which extends south into the 

 Ware River Valley. An inspection of the map will show that this cliamicl 

 is the direct continuation of the Orange-Enfield valley, and it is probal)]e 

 that this longitudinal valley was deeply excavated in pre-Glacial times 

 along the course of Beaver Brook, the north-south portion of the Ware 

 River, and southward jjast Calkins Pond and the Monsou Valley into the 

 Willimansett Vallev. 



The effect of the ice in this position is further seen in tlic iilling of tlie 

 high-lying valley which runs south along the east slope of Quabin Hdl, 

 which rises south of Enfield. 



The sands that pass up the West Brancli of Swift River go Ijeyond its 

 head Avaters and end at Wendell Center, on the divide between the Chicopee 

 and Slillers river drainages, in an instructive way. (See p. 57-1.) The 

 broader sands of the Enfield basin go across Hampshire County into Frank- 

 lin County and end at Orange, as far north as, but mucli lower than, the 

 West Branch sands. 



On the map I have given a single color (1 s) to all the sands here dis- 

 cussed, because they form a continuous series with uniform slopes and 

 because they were plainly formed by a continuous series of events wliicli 

 can not in mapping be conveniently subdivided. It is, however, jjrobable 

 that these two valleys — the West Branch and the Orange-Enfield — remained 

 main drainage arteries until ^Millers River, far in the north, was set free of 

 ice, and thus long after the Ellis ]\Iills outlet was exchanged for the perma- 

 nent Chicopee River outlet. 



It seems, however, that the broad basin was in a sense filled stepwise 

 from south to nortli, so that when the outlet was transferred from the Ellis 

 Mills Valley to the Chicopee River Valley the Swift River began to cut 

 through the sands already deposited and the upper portion of the Vjusin. 

 remained still a catchment area for the flood waters. I deduce this from 

 the fact that the sands of the Belchertown plains, which have clear relation 

 to the present drainage outlet at Three Rivers, are derived entirely from 



