LATER MORAINES OF LAKE MICHIGAN, SAGINAW, AND HURON-ERIE LOBES. 289 



the maximum of the ice, its south-central part was uncovered and became a flat insular area 

 in the ice field, while the two lobes were still intimately united for nearly 200 miles southwest 

 from London. This is called Ontario Island and was at first about 70 or 80 miles long and 

 10 to 20 miles wide. 1 It seems certain that at this stage of retreat the two lobes, though still 

 united, were differentiated by a depression or crease on the surface of the ice along the line 

 of contact. If any drainage escaped from the area of the island it must have followed this 

 crease southwestward, and if the river thus produced carried sediment it would naturally be 

 deposited where the crease came out to the edge of the lobe. 



In Fulton County, Ohio, sand is distributed over a large part of the Defiance moraine. 

 This deposit may have been made by the river just noted, and if there was any tendency to 

 subglacial concentration of drift along its line the result may have been a bowlder belt like those 

 here discussed. The relations observed may be accidental, but the Grosse Isle bowlder belt 

 seems to lead up the slope directly to the sand deposit in Fulton County and its course pro- 

 duced toward the northeast carries it back toward London. 2 The relations of the belts, however, 

 have not been fully worked out and their origin is still problematic. 3 



GROSSE ISLE MORAINE. 



On the central part of Grosse Isle the till is thicker and stands slightly higher than at 

 either its northern or southern ends and it is also somewhat more bowldery than the general 

 surface of the region. The same bowldery drift belt extends a few miles southwest from Trenton 

 and seems to find continuation in the bowlder belt described by Sherzer. This part is also 

 marked by well-defined undulations of the surface which are possibly original morainic features, 

 as maintained by Sherzer, but which seem to the writer to be due largely to the scour of early 

 temporary distributaries of Detroit River. (See pp. 487-489.) 



Where best developed, southwest of Trenton, the ridges are transverse to the course of 

 the moraine and in the whole area the troughs follow the general slope of the surface, curving 

 around from southwestward courses northwest of Trenton to southeastward courses southwest 

 of Trenton. On the Canadian side a well-defined bowlder belt runs west from the high knoll 

 west of Leamington, passing north of Kingsville and Harrow and trending toward Amherst- 

 burg; but it has not been traced across the intervening space. The general relations of the 

 topography to the basin of Lake Erie from which the ice came suggest that this is the normal 

 course of the ice margin from Grosse Isle eastward. 



PKE-WISCONSIN TILL. 



Till older than that deposited by the Wisconsin ice sheet seems to underlie more or less con- 

 tinuously all of the later or Wisconsin drift in Indiana and the southern peninsula of Michigan. 

 It is generally darker in color than the newer drift, is more stony, and is considerably indurated. 

 It is the hardpan of the well drillers and in places is overlain by an old sod bed or by a bed of 

 decayed ferruginous gravel. In typical occurrences it shows much oxidation along its cracks 

 and joints, the oxidized part appearing as bands of brown next to the cracks and in places 

 extending many feet down into the deposit. All of these characters are not likely to be seen 

 at any one locality, however, for at many places the Wisconsin ice sheet scoured away the old 

 soil and the upper part of the older deposit; and the part remaining may be only the deeper 

 part and may show little or no oxidation along its joints. The dark color and the hardness, how- 

 ever, are generally characteristic. The greater number of stones and their more general striation 

 also serve to identify the older till. 



iThe development of Ontario Island is described in a paper by the writer entitled "The moraine systems of southwestern Ontario" 

 (Trans. Canadian Inst., vol. 10, 1913, pp. 1-23). 



2 Sherzer's statement (Geological report on Wayne County, footnote at bottom of p. 82) that this explanation is suggested by the writer for 

 all of the bowlder belts arises from a misunderstanding. The explanation was Intended to apply only to the belt which extends southwest 

 from Grosse Isle toward the great sand deposit in Fulton County, Ohio. 



3 An abstract of a paper on the Crease River was given by the writer before the Association of American Geographers at Baltimore in Jan- 

 uary, 1910. 



34407°— 15 19 



