604 THE GEOLOGY OF MINNESOTA. 



[Channels through the outer moraine. 



a few inches to one or two feet above the water. Lake Shaokatan is about three miles long, and 

 from an eighth to three-fourths of a mile wide, its maximum width being near the middle. Its 

 highest stage is some four feet above the lowest. The southwest end of this lake is at the north- 

 east edge of the morainic belt. Lake Hendricks is three and a half miles long, and its width 

 varies from one-fourth to three-fourths of a mile, being greatest near its northeast end. The 

 maximum depth of each of these lakes is reported to be about 15 feet; and they are bordered on 

 all sides excepting the west by smoothly undulating till, which varies from 10 to 30 feet, or rarely 

 50 feet, above them. Thus the hollows in which they lie sink about 40 feet below the general level 

 of the drift-sheet at the east side of the morainic range, and 25 or 30 feet below the highest 

 part of these channels which are continuations from them through the moraine and the thick sheet 

 of till at its west side. 



Nowhere else for at least fifty miles next to the northwest from Murray county is this mass- 

 ive ridge intersected by any similar channel, and its altitude throughout this distance is from 100 

 to 200 feet above these lakes,. Its highest portion, forming a belt about two miles wide, marked 

 by many hillocBs and hollows, appears to have been pushed out at the margin of an ice-sheet that 

 lay upon its northeast side. The excavation of these channels took place at the same time with 

 the accumulation of this moraine, or more probably at the close of this part of the last glacial 

 epoch, when the ice was being rapidly melted, but before it had receded to its inner line of mo- 

 raine; for the thick mass of the ice-sheet, rising high above its terminal deposits, is the only bar- 

 rier that we can suppose to have existed to turn the course of drainage across this highland, which 

 is now the water-shed between the much lower basins of the Minnesota and Big Sioux rivers, and 

 after this was withdrawn to its later limits at its inner moraine, extending from Spirit lake to 

 lake Shetek and Gary, a lower avenue was opened southward to the Little Sioux river. Without 

 reference to this barrier, it is evident that the course of the waters that eroded these valleys was 

 southwest, because of their extent and fall in this direction. The channel that reaches south from 

 lake Hendricks and then southwest, descends from the summit, one and a half miles south of the 

 lake, with a very gradual slope which probably amounts to 75 or 100 feet in the next ten miles, its 

 width continuing nearly the same as where it intersects the moraine. Another proof that the 

 course of drainage was southwest is the confluence in this direction of the three valleys that 

 cross this range at lake Shaokatou, three miles farther northwest, and at lake Hendricks. On the 

 other side of the moraine no well marked valleys extend northeastward from these lakes; and their 

 outlets, which run only at unusually wet seasons, are turned in a meandering course by slight 

 undulations of the surface. 



There seems to be no indication that the channels through the moraine have been partially 

 filled since their excavation, raising them to their summits, ten to fifteen feet above lakes Benton, 

 Shaokatan, and Hendricks; while yet the position and form of these lakes demonstrate that the 

 portions of the drift-sheet which would have filled their depressions, were carried away by the 

 rivers that cut these gaps. Now it is clear that the overflow from a lake lying between the ice- 

 sheet and its moraine could not excavate a hollow several miles long below a summit which it 

 afterward crossed. Respecting the possible action of subglacial rivers we have little knowledge, 

 but it appears improbable that they could erode such hollows, carrying the material forward through 

 higher channels. It is, however, nearly certain that this removal of the drift belonging upon the 

 areas occupied by these lakes took place while the ice-sheet still covered these areas and reached 

 to its terminal moraine; but near the end of this time, when a warmer climate was rapidly melt- 

 ing its surface every summer, pouring down large rivers to its margin. By such melting the drift 

 which had been gathered into the ice-mass would become exposed upon its surface, and in and 

 near its principal avenues of drainage would be washed away. Only in this manner could the 

 material of the drift-sheet corresponding to the depressions of these lakes be removed by the usual 

 agency, that is, by the current of descending streams. If this be the true explanation, it involves 

 a very important conclusion respecting the amount of drift contained in the ice-sheet and finally 

 exposed by the melting of its surface. Modified drift and kames, as also certain features of the 

 till and of the terminal moraines, prove that the ice of the glacial period became considerably 

 filled with the material of the drift, gathered up into its mass from the land over which it moved. 

 This explanation of the origin of these lake basins indicates that the ice-held drift here amounted 

 to a sheet at least forty feet thick; but much of it may have been in the lower two hundred feet of 

 the ice, below the top of its terminal moraine. 



