320 PLEISTOCENE OF INDIANA AND MICHIGAN. 



even the high lands between them. As it retreated, however, the relative importance of the 

 lake basins in controlling the ice flow increased rapidly and by the time the ice front had with- 

 drawn to the most southerly points of the watershed of the lake basins it had taken on lobate 

 forms of a most pronounced type. In the latest or Wisconsin stage of glaciation the farthest 

 extension of the ice did not reach so far south in the region west of central Ohio as it had before. 

 In Illinois it reached only about halfway from the shore of Lake Michigan to Mississippi 

 and Ohio rivers. As the ice drew back, each lake basin, at the time of most pronounced 

 lobation, had its ice lobe, which conformed to the outlines of its southern part with remarkable 

 fidelity. The moraines laid down along the margin of the lobes at this stage are roughly con- 

 centric with the basins and nearly parallel with their southern shores. 



OSCILLATIONS OF THE ICE FRONT. 



The periodic oscillations in the retreat of the Wisconsin ice sheet introduced a peculiar 

 complexity into the lake history. It has been shown above that the stronger or principal 

 moraines formed at the culmination of a readvance mark relatively long intervals of time — 

 several hundreds of years, perhaps more than a thousand. The readvance to the moraine 

 covered a very considerable distance, certainly in some cases 25 to 50 miles and perhaps twice 

 as much. The long intervals of time between a moraine that precedes and one that follows 

 such an oscillation accounts perhaps to some extent for the commonly observed discordant 

 relation between them. 



Important oscillatory variations in the retreat of the ice produced a corresponding effect 

 on the lake history, for after lakes were lowered by a retreat of the ice the readvance was likely 

 to close their outlets and raise their waters to higher levels. Three times, certainly, and per- 

 haps four or five times such a change affected the waters of the Huron-Erie-Ontario basin. 

 During several oscillations the ice front stood in critical relations to the land barriers that held 

 up the lakes and relatively slight changes in position opened or closed outlets and changed the 

 level of the lake waters. 



In view of the apparent character of the glacial oscillations it seems necessary to take 

 account also of the halts of the ice front at the back steps or climaxes of retreat. These probably 

 affected the lakes as greatly and lasted as long as did the halts on the moraines which mark 

 the climaxes of advance. In one place at least there is remarkably clear and complete proof 

 of the long duration of a particular lake stage (Arkona) winch existed during the pause at a 

 back step or climax of retreat. The changes of the lakes during the retreat of the ice front across 

 the Huron-Erie-Ontario basin are the most complicated now known and form the central theme 

 of the discussion that follows. 



THE FIRST LAKES. 



As soon as the ice front withdrew to the north side of the southern watershed of the Great 

 Lakes small ice-dammed lakes began to be formed. In Ohio a number of such lakes appeared 

 along the north side of the divide south of Lake Erie. It seems a necessary inference that in 

 consequence of the oscillations of the ice many of these earliest small lakes were formed at the 

 extreme position of retreat and were overridden and obliterated by the next readvance. Indeed, 

 some of them may have been formed and overwhelmed two or three times before the larger lakes 

 became permanently established. There is some evidence of this sort of development at Fort 

 Wayne, Ind. 



The earliest small lakes discharged at first southward independently through several gaps in 

 the divide ; a little later they fell to lower levels and discharged westward to a lower gap ; and finally 

 they discharged into the first small, narrow representative of glacial Lake Maumee. Mr. Lev- 

 erett has already described these earliest lakes in Ohio; 1 he has also described others of some- 



i Leverett, Frank, Glacial formations and drainage features of the Erie and Ohio basins: Mon. U. S. Geol. Survey, vol. 41, 1902, pp. 610-611. 



