CUESTERVILLE-LEEDS SYSTEM. 197 



small kames near the jaws of low passes. The great size of the gravels in 

 Chesterville demands a large supply of water from the north. For these 

 reasons I consider it highly probable that the gravels of the upper valley 

 of the Sandy Eiver are partly an osar-plain and ])artly an overwash or 

 frontal plain, and that this glacial river drained a large area north of Phillips 

 and sonth of Mount Abraham. From Farmington Falls south the probable 

 course of the glacial river Avas along the valley of Chesterville Stream. 

 The relations of this osar' system to sedimentary clay and sand are inter- 

 esting. From Chesterville south this system is, throughout its whole course, 

 flanked and partly or wholly covered by a broad plain of sedimentary, 

 bluish-gray clay, overlain by more or less sand. Toward the north this 

 clay plain connects with the similar plain found in the valley of the Sandy 

 River by two low valleys, one along the Chesterville Stream and the other 

 lying 2 or 3 miles east of it. The broad Chesterville plain of sedimentary 

 clay connects with a similar plain that borders the Andi'oscoggin River by 

 two routes, one around the northern base of Moose Hill, in Jay, and the 

 other along a low pass that leads northwest from near the Camp Grround in 

 East Livermore. Whether the water flowed from the Sandy River over 

 into the Androscoggin or in the opposite direction is uncertain; possibly the 

 flow was alternately in opposite directions, as the flood height of these 

 rivers varied. South of the Camp Ground the clay plain bordering the 

 osar is continuous with that of the Androscoggin Valley, as far as North 

 Leeds, where a hill intervenes between the two plains. South of this point 

 we have, in addition to the Andi-oscoggin plain, two other plains covered by 

 clay. One lies directlj^ along the line of the osar, past Curtis Corner to 

 Leeds Junction and Sabatis Pond. Another is from 1 to 3 miles west of 

 the last named and occupies the eastern base of Quaker Ridge in Grreene. 

 A short distance north of Greene station this plain turns east to the head of 

 Sabatis Pond. All of the clay plains just described are above the contour 

 of 230 feet except at their south ends, near Sabatis Pond and Lewiston, and 

 near Andi'oscoggin Pond. Wherever I crossed them they filled the valleys 

 they occupied from side to side, as if they were valley alluvium. On gen- 

 eral grounds we might expect the deposition of osar border clays in a l:)road 

 ice channel along the flanks of the gravels, but if such were deposited they 

 seem to be lost in the midst of the fluviatile clays and sands that were 

 deposited later. It will require some nice discriminations in order to mark 



