Eiiglacial Drift in the Mississippi Basin. — Upham. 371 
red sand and gravel owe their color to red sandstones, slijiles 
and conglomerates, of the Keweenawan series in northern 
Wisconsin and northeastern Minnesota, whence the ice-sheet 
flowed southwesterly over a large part of eastern Minnesota, 
continuing southward on the west side of the Wisconsin drift- 
less area. That early glaciation doubtless was confluent with 
another glacial outflow from the region of lake Winnipeg 
and Reindeer lake, carrying boulders of limestone fiom Mani- 
toba southward over the Dakotas and the greater part of Alin- 
nesota. South of the driftless area the severed parts of the 
ice-sheet reunited, and its utmost limit crossed southern Ohio. 
Indiana and Illinois, central Alissouri, northeastern Kansas, 
eastern Nebraska and the country adjoining the southwest side 
of the Missouri river in the Dakotas and Montana. From the 
maximum extension thus outlined, the ice-sheet had been melt- 
ed back about four hundred miles, but only about seventy-five 
miles from the nearest northern part of the driftless area, 
when the red modified drift at Rush City was deposited. How 
long that tract remained a land surface we cannot reliably de- 
termine; but not improbably it was so during the long interval 
between the Kansan and lowan stages of glaciation. Five miles 
west of Rush City, wood and peaty matter occur in a layer 
of clay at the top of the modified drift, covered by 8 or 10 
feet of the later northwestern till, showing that after a stage 
of recession the ice-sheet again extended over this sand and 
gravel plain, this time flowing here in a direction nearly at right 
angles with its former course. Contemporaneously, however, 
at no great distance eastward, there was quite surely a conflu- 
ent glaciation from the northeast. Upon a belt reaching from 
St. Paul and Minneapolis northward across Minnesota the 
northeastern and northwestern glacial currents met and con- 
tended, now one and now the other pushing back its oppon- 
ent, during all the successive stages of the Glacial period. 
Let us next consider the thickness of the very uniform 
sheet of superficial till at Rush City in comparison with its 
area. Its linear extension, across the tract where it was as- 
certained by many wells to be continuously underlain by the 
early modified drift, exceeds its thickness more than a thou- 
sandfold. Its mode of deposition, having been englacial and 
afterward superglacial drift, may be best indicated by an ex- 
tract from my report (Minnesota, vol. ii, p. 417), as follows: 
