98 BULLETIN OF THE ESSEX INSTITUTE. 



tend to clog ill the plane between the live and dead ice. 

 There might thus be established one of those masses of 

 till involved in the ice which Chumbeilin has described in 

 Greenland. On the subsequent melting out of the ice, 

 the unequal thickness and rate of lowering of this till to 

 the ground would result in mounds. This hypothesis 

 accounts for the till in the submarginal moraine but does 

 not account for the underlying watervvorn gravels. On 

 this account, the first hypothesis is preferred. 



Mr. H. T. Burr, a student in Harvard University, has 

 traced this line of ice-front several miles to the northeast. 



The Wrentham- Weymoutli line of lakes, — There is 

 a prominent line of glacial lakes extending in a north- 

 east aud southwest direction from near the northeast 

 corner of Rhode Island to Weymouth, Mass. These 

 lakes are as follows, begiiming on the southwest : Shep- 

 ardville Reservoir, Shepard's Pond, Cocasset Pond, Ne- 

 ponset Reservoir, Billings Pond, Massapoag Pond, ponds 

 and reservoirs at Canton, Ponkapoag Pond, Great Pond, 

 Little Pond. These lakelets are simply the water occu- 

 pied portions of low areas partly surrounded by plains of 

 sand and gravel. No attempt has been made to map this 

 line of apparent ice-front and further study is necessary 

 to show that the plains are not merely fans fringing ice- 

 blocks. 



The enclosing plains form a line of deposition not 

 readily separated from the wash-plains referred to in this 

 paper as the Woonsocket-Sharon line described below. 

 By the frequency of the three-hundred feet level on some 

 of these deposits from East Foxboro northward towards 

 Sharon, it seems probable that further study will show a 

 connection between the plains de[)endont on water-level 

 in this field. 



The Woonsocket-Sharon line of dejjosits. — A fairly 



