GLACIAL LAKE SAGINAW. 361 



character it is evident that the water stood at their level much longer than it did anywhere 

 else. This might be expected from the fact that these beaches represent in time the duration 

 of Lake Maumee after the opening of the outlet near Imlay and the whole life of Lake Arkona 

 and probably that of Lake Whittlesey also. They enter the head of the deeper channel 

 at Maple Rapids at an altitude of 35 to 40 feet above its floor. It seems certain, therefore, 

 that the most prominent terrace is the old floor of the channel during the times of Lakes 

 Arkona and Saginaw — the strongest beach corresponding in time to the strongest terrace. 

 Then sometime later, probably coincident with a drop in the level of Lake Chicago, the river 

 began cutting back from a lower level at its mouth and the old floor was cut away, leaving 

 fragments as terraces along the sides. By the strength of their development and their bowldery, 

 stony surfaces these fragments indicate a relatively old and mature river bed. 



A further evidence of the relatively long duration of the outlet with its floor at the terrace 

 level is seen in the relation of the larger tributary streams which enter the channel from the 

 sides. Flat River at Lowell, Grand River at Lyons, and several smaller streams came in more 

 evenly at grade at the terrace level than they do at the present level. Stony and Prairie creeks 

 show this relation as well if not better than the large streams. Further evidence is found in 

 well-defined terraces of the same floor on Grand River above Lyons. It might be thought that 

 these terraces were formed by the deepening of Grand River since the great outlet river aban- 

 doned Grand River channel. Grand River is perhaps large enough to give a slender basis of 

 plausibility to this idea, but the present condition of the channel floor shows that Grand River 

 has done virtually no deepening in the channel, except in its own actual bed. Outside of this 

 it has slightly built up the floor by the deposition of silt. Maple River is so very small and feeble 

 that it has accomplished virtually no deepening. 



RELATION OF ARKONA BEACHES TO LAKE SAGINAW. 



The relation of the present head of the channel to the Arkona beaches is somewhat peculiar 

 and significant. The head is not in the deeply trenched part of the valley but is in a swamp 

 on a wide, flat clay plain sloping gently toward Saginaw Bay, 16 or 17 miles east of Maple Rapids. 

 The divide, which is in a depression so slight that it is scarcely perceptible, is about 58 feet 

 below the level of the upper Arkona beach 2 miles northeast of Elsie. 



These relations show clearly that both the original divide and the divide that held so long 

 and changed so little during Lake Arkona were at a considerably higher level. But the divide 

 at the time could not have been higher on the plain where it is now. To be higher it must 

 have been a number of miles farther west — at least as far west as the moraines that cross the 

 river 1 to 3 miles east of Maple Rapids. 



At Ithaca the old lake plain slopes gently eastward from the inner base of the moraine, and 

 there is much reason to believe that it did so originally from the east front of the Flint moraine 

 east of Maple Rapids. As the head of the channel receded eastward it was therefore gradually 

 lowered as its place moved down the eastward sloping plain. But with the floor of the channel 

 heavily protected by bowlders, the lowering was slow and required a long time. Only in some 

 such way is it possible to explain the present height of the upper Arkona beach above the divide. 



