it is next to be noted that the small lateral streams which now enter the old river course find it much too large for their volume. This is especially true near the divide across which the old channel was cut. It is only at some distance down the channel that enough water has entered through lateral streams to form a considerable river; yet all along the channel maintains about the same width. Evidently, therefore, it was not cut out by the existing drainage. In consequence of the small volume of the longitudinal streams now occupying old channels, they are frequently more or less obstructed by the alluvial fans built at the entrance of lateral tributaries; thus swamps or long, narrow, lakelike expansions of the rivers are produced up stream from the fans. This was first noticed by Warren, and since then the list of examples has been greatly increased by Upham and others. In the last of Upham's papers referred to above, he describes a number of lakes of this kind on the Qu'appelle and Pembina Rivers and elsewhere. Long Lake in Assiniboia is about fifty miles in length, but only one or two miles wide. Lac qui Parle and Lake Traverse, in the old channel at the head of the Minnesota River, are of this kind. The sluggishness of the Minnesota and of the Illinois Rivers just above their junctions with the Mississippi has been attributed to the same cause, and this would indicate that at the time when the channels of the Minnesota and the Illinois were occupied by the large rivers which once flowed through them, these held the place of main streams, while the Mississippi came into them with smaller volume as a tributary.
There is on this account a curious contrast to be noted between the excavation of the late glacial channels that were cut out during the closing stages of the Glacial period by the overflow from glacial lakes, and the clogging of the preglacial valleys that were commonly filled with sands and gravels by streams that came directly from the retreating ice front without delay and filtering in lakes, as in southeastern Ohio. Both of these kinds of valleys mark the courses of "constrained" drainage near the end of the Glacial period. At the ice front the water supply in both cases was doubtless surcharged with detritus; but the waters that had to accumulate in lakes marginal to the ice front before flowing away as rivers must have been nicely filtered, so that they issued clear and blue from the lake outlets; while the others had to carry their detritus down stream for many miles, and must have been of gray and turbid color for long distances. The plentiful clear waters of the streams of the first class ran down the valleys that led from the lowest pass in the lake rims and cut down their channels to a moderate grade, oftentimes so moderate that the present river occupants of the valley are unable to keep them clear of the alluvium that is brought in by tributaries;
