26 PROCEEDINGS OP BROOKLYN MEETING. 



and principal stage of lake Algonquin is shown by these beaches to have coincided 

 closely in area with lakes Michigan and Superior, but to have been considerably 

 more extensive eastward than the present lake Huron and Georgian bay. It held 

 a level which now by subsequent differential epeirogenic movements is left prob- 

 ably wholly below the level of lake Michigan by a vertical amount ranging from 

 almost nothing to about 40 feet. Its shores were nearly coincident with the west- 

 ern shore of lake Huron, but eastward they are now elevated mostly 150 to 200 feet 

 above that lake and Georgian bay, and in the lake Superior basin they vary from 

 about 50 feet above lake Superior at its mouth and along its northeastern and 

 northern shores to 25 feet at Houghton, and to a few feet or none at Duluth. The 

 earliest outflow of lake Algonquin appears to have passed southward bj^ the present 

 course of the Saint Clair and Detroit rivers; thence it ran east as a glacial river 

 Erie, following the lower part of the bed of the present lake Erie, which then had 

 an eastward descent of probably 200 feet, allowing no lake or only a very small one 

 to exist in the deepest depression of the basin ; and north of Buffalo it coincided 

 with the course of the Niagara river. 



Order of Kecession of the Ice-sheet territorially 



The recession of the ice-sheet from the area of lakes Warren and Algonquin was 

 earlier than from the lake Ontario or Iroquois basin and the country eastward to 

 the New England coast. 



During the time of formation of the high Belmore and Nelson beaches of lake 

 Warren, this glaciallake, outflowing at Chicago, stretched northward to the north 

 side of lake Superior and northeastward to lake Nipissing. With lake Warren of 

 this extent, the ice-sheet had melted off from all the northern United States west 

 of lake Nipissing and of Buffalo, New York ; but yet, to form a barrier on the east, 

 it remained u nmelted upon all the Niagara and lake Ontario or Iroquois area. Thus 

 we see that all the moraines within the limits of the United States west of the 

 great angle of the drift boundary near Salamanca, in southwestern New York,* are 

 somewhat older than the moraines east of that angle in New York, Pennsylvania, 

 New Jersey, Long island and New England. This difference in age, however, be- 

 tween the western and eastern moraines and drift was perhaps no more than 500 

 to 1,000 years, as we may infer from the rate of retreat of the portion of the ice- 

 front forming the northern barrier of the glacial lake Agassiz. 



This unexpected view of the order of departure of the ice-sheet finds meteoro- 

 logic explanation as follows: The melting of the vast western part of the ice-sheet 

 in the United States, from North Dakota and Minnesota east to the lake Erie basin, 

 would supply to our eastwardly moving storms a very great amount of moisture to 

 be precipitated farther east. That precipitation I think to have been mainly snow, 

 as these storms, moisture-laden from the western ice-melting, swept over the more 

 eastern part of the ice-sheet. Hence the eastern great ice-lobe from Salamanca to 

 Long island. Cape Cod, and the gulf of Maine, would be fed and thickened and 

 spread in some places even beyond its previous limits, while all of the ice-sheet 

 farther west in the United States was being melted away, t 



* Consult Professor Chamberlin's maps of the glaciated areas of the United States, U. S. Geol. 

 Survey, Third Annual Report, plates xxviii and xxxiii ; Seventh An. Rep., plate viii. 



t For evidence of similar but smaller climatic effects on the waning ice-sheet in Minnesota see 

 Proc. Am. Assoc. Adv. Sci., vol. xxxii, for 1883, pp. 231-234, and Geology of Minnesota, vol. ii, 1888, 

 pp. 254-256, 409-413. 



