COKEELATIVES OF BLOOMINGTON MOBAINIC SYSTEM. 109 



Brandywine Creek the topography is of a subdued swell and sag type. Its depression below the 

 bordering uplands for most of its length is only 15 to 20 feet, but this does not represent its 

 original depth, for it is deeply filled with peaty material which caused much trouble in the 

 early construction of roads across it. 



The range in the altitude of this channel is not great, and the divides crossed by it are of 

 remarkably similar heights, as shown in the table below. Beyond the last divide in central 

 Hancock County the altitude gradually decreases down the channel. 



Range in altitude along abandoned channel in Grant, Madison, Hancock, and Shelby counties, Ind. 



Summit on divide between Mississinawa River and Pipe Creek near Grant-Madison county 



line (canal survey, 1835) 8g2 



Pipe Creek valley at Alexandria (Lake Erie & Western Railroad survey) S50 



Summit on divide 3 miles south of Alexandria (barometric) 890 



White River at Anderson, 25 to 30 feet above present bed (canal survey) 850 



Summit south of Anderson (canal survey) 889 



Pall Creek valley (Cleveland, Cincinnati, Chicago & St. Louis Railway survey) 845 



Summit between Sugar Creek and Brandywine Creek (Cleveland, Cincinnati, Chicago & St. 



Louis Railway survey) 885 



Near Greenfield (Pittsburgh, Cincinnati, Chicago & St. Louis Railway survey) 835 



Origin. — Two hypotheses as to the origin of such a channel may be considered. By the 

 first it is assumed that the summits noted above mark successive positions of the ice margin 

 and that the portion of the channel between each summit and the stream south of it was occupied 

 by a stream which flowed no farther south but turned down the intersecting valley. For 

 example, the broad channel leading from the summit near White River, just south of Anderson 

 southward to Fall Creek would be the product of a later stream than the channel leading from 

 the Fall Creek and Sugar Creek divide southward to Sugar Creek, and stdl later than the one 

 from the Sugar Creek and Brandywine Creek divide to Brandywine Creek, but earlier than the 

 channel north of White River leading from the divide south of Pipe Creek to White River. 

 Under this hypothesis the ice may be conceived to have had either of two quite different posi- 

 tions relative to such portions of the channel as lie between a given divide and stream. At the 

 time of the formation of each channel the ice may have had its margin at the southern end 

 of the channel, in which case the channel would have been formed by a subglacial stream; or 

 it may have had its margin at the summit, in which case the channel would have been formed 

 or at least enlarged outside of the ice sheet by waters that were not confined by ice walls. Cer- 

 tain phenomena, however, seem to be incongruous with this latter conception; for instance, 

 the Anderson esker, which occupies the channel and rises in places higher than the bluffs of 

 the channel and yet was evidently formed in it after the channel had been excavated, seems 

 to call for the presence of the ice sheet and the confinement of ice walls; and again, certain 

 gravel deposits that fringe the channel and rise in places somewhat above the level of the bor- 

 dering till tracts seem to call for a stream that followed approximately the course of the channel. 

 This second conception, therefore, seems untenable at least for that part of the channel, and 

 the conception of a subglacial origin for the channel seems more nearly in accord with the 

 features along the valley. 



The second hypothesis as to the origin of the channel starts with the assumption that the 

 channel was formed from end to end by a single subglacial stream that started some 60 miles 

 back from the margin of the ice and flowed to the margin, rising and falling with the rise and 

 fall of the surface of the ground beneath the ice sheet but ever keeping straight onward because 

 the superincumbent ice sheet prevented escape to either side. The hydrostatic pressure is 

 assumed to have been strong enough to have carried the stream over the low summits which 

 lay in its course. The fact that the channel is not disjointed between the sections, as assumed 

 in the first hypothesis, and the further fact that it rises from the valleys southward to the divides, 

 maintaining similar sharply defined bluffs up the slope, across, and down the slope of each 

 divide, lend much strength to this hypothesis. Possibly there was an initial opening of the 



