428 J. D. Dana — xSW/// //"///•// T>>.sr}ni,(j, of 1 J <}.■<■ W'aiv'q>"J- 



Station eight is on the east side of the river, and one-half 

 mile below South Win in, and also at station 



nine and those above on the Natchaug, the floods were gener- 

 ally very widi h -piead, forming oivat areas of level plains 

 which extend far hack from the streams, even to Mansfield 

 Center, four miles north of the city of Willimantic. 



Unlike the region just mentioned, the Willimantic River lies 

 in a very narrow valley with the lines of hills on either side 

 close to the stream, so that the waters swept everything out of 

 the valley, and nothing but doubtful remains of high terrace- 

 deposits are found till we get to Eagleville, seven miles above 

 it* mouth, and at but few points above this were we able to 

 determine the height of the floods. 



Agricultural School, Mansfield, Conn., Oct. 7th, 1882. 



The most remarkable of the changes that are known to have 

 orenrred in the water-courses of North America is that in the 

 discharge of Lake Winnipeg from a former southward course by 

 the Minnesota Channel and Mississippi to its modern discharge 

 into Hudson's Bav, first announce! and sustained by General G. 

 K. Warren, in a report of 18*7 published in the Report of the 

 U. S Engineers for the year 1868 (pp. 307-314), after levelings 

 along these rivers, by order of the government, in 1866 and 

 1867. The question was more fully illustrated by General 

 Warren in "an Essay concerning important physical features 

 exhibited in the Valley of the Minnesota River and upon their 

 signification," submitted to the Chief of Engineers in 1874 and 

 published in the Report for 1875 (pp. 385-402); and afterward, 

 further discussed bv him in his paper <m the I.ri- giniz oi the 

 Upper Mississippi, in the Report for 1878 (pp. 909 to 9l>6) with 

 a reproduction of some of the maps n{ the essay of 1874. The 



as to seeli- • r,fin tl,is'~lournal 



on pages 417 f-'U. vol. xvi, 1878, along with its eight maps 

 and plates. 



In the first of his papers, that of January, 1867, General 

 Warren, after mentioning the evidences that "Lake Winnipeg 

 was once" continuous southward over the central portion of the 

 Red River of the North, and had its outlet down the Minnesota, 

 and not down the Nelson to Hudson's Bay" (p. 307), considers 

 the origin of the form ei hvdrograpim d conditions. He speaks 

 of the possibility of an ice-barrier on the north in the Glacial 



