CHESTEEFIELD-LBEDS SYSTEM. 199 



one place this gravel has been excavated by the Maine Central Railroad 

 Company. There is an interval of at least 3 miles between this ridge and 

 the delta-plain at Curtis Corner, which forms the apparent termination of 

 the Chesterville-Leeds system. 



About 4 miles soiitheast of Leeds Junction a large mound rises in the 

 midst of the large swamp at the north end of Sabatis Pond. It is probably 

 composed of glacial gravel. 



At various points along the shores of Sabatis Lake there are small bars 

 and terraces of glacial gravel at various heights aboA^e the lake up to 100 

 feet. The material is but little waterworn and forms a thin cap of semi- 

 morainal yet water- washed gravel overlying the till. ' It is uncertain whether 

 these gravels south of Curtis Corner are any part of the Chesterville sys- 

 tem. I provisionally marked them as distinct. The gravels along Sabatis 

 Lake, taken in connection with the terminal moraine at Sabatisville, afford 

 some prima facie evidence of a local glacier moving down the valley of 

 Sabatis Lake, which is boi'dered by hills several hundred feet high. The 

 shortness of the moraine shows that the ice movement was then confined to 

 the valley. North of Sabatis Pond are two open valleys, along which the 

 ice could easily flow on a descending grade to Sabatisville. One opens 

 northward into Monmouth, the other extends northwestward through Leeds 

 toward Wayne and East Livermore. After the general movement of the ice- 

 sheet had ceased, on account of transverse hills, ice could still for a time con- 

 tinue to flow in these favorable valleys. Such a local tongue of ice in the 

 valley of Sabatis Lake would account for: (1) the terminal moarine at Sabat- 

 isville; (2) the water- washed moraine stuff on the sides of the hills near the 

 lake (i. e., these were formed along the margin of the local glacier); (3) the 

 fact that the basin of the lake was not filled up by the clays, which may be 

 due in part to the fact that the valley was filled by ice till a rather late date. 

 The length of the system from Chesterville to Curtis Corner, Leeds, is 20 

 miles. This portion of the system must date from late glacial time. 



I have not here explicitly classified the water drift of the Sandy River 

 above Farmington Falls as an osar-plain overlain by later frontal sediments. 

 The critical reader, however, who compares this system with those of the 

 other valleys lying eastward at the same distance from the coast, as, for 

 instance, the gravels of the Carrabassett and upper Kennebec valleys, will 

 discern that the sedimentary drift of all these valleys has many features in 



