W. Ujpham — Diversity of the Glacial Drift. 363 



uneven distribution. The early accumulation and advance of 

 the ice to its extreme limits gave a comparatively thin ice-sheet 

 with feeble erosive action on all the outer part of the drift- 

 bearing area. Its drift there was nearly all brought from con- 

 siderable distances at the north and was deposited in obedience 

 to the glacial currents of the marginal portions of the ice- 

 sheet. Now we have upon many districts of the thick later 

 drift the remarkable aggregations of the till called drumlins, 

 which appear to have been amassed by convergent currents of 

 the ice-sheet during its retreat.* Similar selective action of 

 the outflowing early ice advance close to its farthest limits I 

 think to have amassed that outermost early till in the patches 

 where it is now found, having received little change by later 

 erosion. 



The well oxidized and leached condition of the early outer 

 drift everywhere is easily referred to its derivation chiefly 

 from the preglacial residuary clays, decaying rocks, weathered 

 rock cliffs and tors, and bowlders of secular disintegration. 

 Again, its smoothed surface, without the inequalities of accu- 

 mulation which provide basins for the myriad lakes and lake- 

 lets of the later drift, seems attributable to the gentle currents 

 of the early thin ice-sheet, in contrast with which the late 

 thick ice powerfully eroded its rock bed, even close to the 

 boundary, and tumultuously heaped or very irregularly spread 

 its drift with many lake-enclosing hollows. 



Forest beds between deposits of till in northeastern Iowa, 

 and in portions of other states of the Mississippi basin, testify 

 of glacial recessions and re-advances, the ice-sheet probably 

 recovering at some times marginal belts from 50 to 100 or 

 200 miles wide of its previously lost ground. These great 

 oscillations, however, need not have required a very long time, 

 certainly no more than a few thousand years for them all, as 

 we may well learn from Prof. I. C. Russell's observations of 

 the drift-enveloped and forest-clad borders of the Malaspina 

 ice-sheet between Mt. St. Elias and the ocean. 



Amid the waverings of the retreating ice, often large chan- 

 nels were cut in the early drift and became covered and par- 

 tially filled by the later drift. In southern Minnesota these 

 old water-courses are recognized as far northward as the Min- 

 nesota river valley, considerably to the north of all our re- 

 corded observations of forest and peat beds enclosed in the 

 till sheet. They seem to have been probably rapidly eroded 

 when the altitude of the country and its slopes of descent 

 from north to south were greater than now. The few thou- 

 sand years which are here regarded as the time of the fluctua- 



* "Conditions of Accumulation of Drumlins," Am. Geologist, vol. x, pp. 339- 

 362, Dec, 189-2. 



