TOPOGRAPHIC EXPRESSION OF THE DRIFT BORDER. 39 



seldom, if ever, aggregated in the form of knolls or ridges. As noted above, 

 it is not certain but that the deposition was largely made by streams, rather 

 than by direct glacial action. 



Below St. Louis there is a less regular and lighter deposit of drift in 

 the vicinity of the border than in districts to the north, and the border there, 

 so far as noted, is without topographic expression, the drift being found 

 largely in depressions, with only a thin veneering on the hills. As noted 

 farther on, a prominent belt of drift ridges comes down nearly to the drift 

 border from the northeast across southeastern Madison, central St. Clair, 

 eastern Monroe, and northern Randolph counties, and there turns southeast- 

 ward, taking a course nearly parallel with the drift border and scarcely 10 

 miles back from it, Upon turning southeastward this belt of ridged drift 

 becomes ill-defined, but has been traced with some certainty to central 

 Jackson County (midway between Ava and Murphysboro). As yet no line 

 of ridges marking a continuation has been found farther southeast. It is 

 possible that the sheet of drift which terminates at this belt of ridges may 

 come to the glacial boundary in southern Illinois, and constitute that 

 boundary from there eastward. 



In southern Illinois occasional low knolls, 20 feet or less in height, 

 occur in the vicinity of the drift border, and there appears to be a slight 

 ridging in east-west direction in the southern portion of Williamson County, 

 a ridging sufficient to influence the course of streams, as indicated on a later 

 page (p. 527). In the vicinity of the Wabash River, near Ridgway, and 

 thence northeastward to New Haven, Illinois, there is a belt of low sandy 

 knolls and ridges, the origin of which is not clearly determined. Possibly 

 they are entirely the result of wind action, or they may be due in part to 

 glacial action. 



In southwestern Indiana a few places were found where the drift 

 border and districts immediately back of it show a tendency to aggregation 

 in low knolls and ridges. Perhaps the most conspicuous instance is found 

 in Gibson County, near Fort Branch, where for a distance of about 3 miles 

 along the east side of the Evansville and Terre Haute Railway there is a 

 ridge of drift 30 to 50 feet in height and nearly a mile in breadth, whose 

 surface is quite undulatory. From the southern end of this ridge south- 

 westward into Posey County knolls 10 to 20 feet in height are of frequent 

 occurrence, and in places become so closely aggregated as to give a 



