LAKE KANKAKEE. 335 



ward retreat of the ice margin. The bowlder belts may mark the successive 

 ice margins. In this case the ridging of the sand may be due entirely to 

 wind action, or the ridges may be esker-like accumulations formed beneath 

 a nearly stagnant ice sheet. Perhaps a systematic relation between the 

 bowlder belts and the sandy areas may yet be worked out. At present the 

 courses of the bowlder belts seem less systematic than would be expected 

 under this hypothesis, so that this matter is still in question. The presence 

 of beds of peat and muck at the base of the sand, which is a conspicuous 

 feature in the vicinity of Reynolds, Indiana, perhaps opposes this interpreta- 

 tion, since it calls for an interval of emergence with exposure to atmospheric 

 action between the withdrawal of the ice sheet and the deposition of the 

 sand. The development of the peaty muck may, however, not prove fatal 

 to this interpretation, since it is apparently limited in extent and demands 

 no great length of time. The outwash, where most vigorous, may have 

 caused the sand to encroach upon peat bogs in the outlying districts. In 

 some cases it is probable that wind may have drifted the sand in such man- 

 ner as to bury the peat bogs. That wind action has been influential in 

 heaping up the sand throughout this area can scarcely be questioned. 



An alternative interpretation to that of a gradually formed glacial out- 

 wash at the margins of receding ice lobes is found in the hypothesis that this 

 sand-covered region is the bottom of an extensive lake in which the waves 

 washed the till sheet and formed the sand. This hypothesis need not require 

 that the entire sandy area was at any one time covered by a lake. It may 

 also admit the application of the hypothesis of glacial outwash in explanation 

 of the gravel and sand on the east and north borders, where the altitude is 

 greater than on the south and west borders of the sand-covered area. It 

 would differ from the former hypothesis in requiring a larger amount of 

 static water and greater influence of waves. Were the upper or southern 

 ridg-e of sand on the south border of the sand-covered region horizontal or 

 even slightly tilted toward the east, it would give support to this hypothesis, 

 for the ridge certainly bears a strong resemblance to a lake beach, but the 

 fluctuations in level (from 675 to 750 feet above tide) seem too great to be 

 accounted for by the warping of an originally horizontal line, and esjjecially 

 since the departure from the horizontal is in the form of wave-like oscil- 

 lations instead of a regular uplift. Furthermore, the nearly parallel ridge 

 a few miles to the north does not show similar warping. The supposed 



