188 



across Pretty Prairie, a gravelly terrace plain, a distauce of three miles 

 to the mouth of the Tippecanoe River. The crest of this moraine at the 

 Soldiers' Home, four miles north of Lafayette, is higher than the surface 

 of the plain at Monticello or Winamac, and the gap has the appear- 

 ance of having been once the passageway for a large stream from the 

 north. The part of the Tippecanoe from New Buffalo to the great bend 

 is the newest part of the stream. It established its meandering course 

 among the sand ridges along the eastern side of the lake bed and con- 

 nected the part above the bend, which formerly flowed into the lake, with 

 the part which was the lake outlet, giving an interesting example of 

 a spliced stream. 



The description of the development of the drainage of the Wabash 

 system has been traced to the above extent in order to group its main 

 facts together and bring them to bear on the question of the manner of 

 recession of the ice sheet from its basin and some of those basins adjoin- 

 ing it. 



Several writers on problems connected with the drift area seem to 

 assume that the ice sheet could not have receded in any other way 

 than from west to east. The Kankakee Lake, the western Indiana bowlder 

 belts and various other problems are perplexing problems on this assump- 

 tion. While in a general way the view is doubtless true that the recession 

 was in this direction, the solution of several interesting points connected 

 with Indiana drainage becomes simple by the acceptance of good evi- 

 dence that in western Indiana the recession was from east to west. 



The Michigan, Huron and Erie depressions were doubtless lines of 

 southward and southwestward movement which became filled with ice 

 and overflowed before the country between was invaded. Gradually the 

 ice accumulated and covered the crests of the divides, becoming a con- 

 fluent area with smooth, regular slopes on the surface, but conforming 

 generally on the under side to the relief of the rock surface below. 

 Vallej's and low tracts of the preglacial surface would become lines of 

 more rapid flow and the ice would move farther foi'ward along these 

 lines than elsewhere. The arrangement of the moraines in Illinois, In- 

 diana and Ohio shows the influence of this lobate movement to the 

 limits of the drift of any period. 



The curving to the north of the glacial boundary in Indiana is easily 

 explained by the stranding of the ice along the north and south belt 

 of resistant rocks, including the Knobstone in that part of the State, while 



