hence they are frequently characterized by swamps or linear lakes, especially near their heads. On the other hand, the streams of the second class, surcharged with detritus from the ice, frequently demanded a steeper grade than they found, and therefore deposited much of their load, filling up their valleys with broad flood plains of sand and gravel, such as are now growing in front of the Greenland and Malaspina glaciers; and sometimes they may have transformed side streams into lakes, such as now characterize the side streams of the Red River of Louisiana.
The record of the latter feature is truly a rarity in the past, but it has been deciphered by Andrews as perceptible in some of the lateral tributaries of the gravel-filled valleys of Ohio. The following is Andrews's account. After describing the terraces of sand and gravel derived from the glacial drift and occurring along those streams whose sources lie in the "area of the general drift," he says: "There is in the second district [southeastern Ohio] another and very distinct system of terraces, found on streams emptying into the larger streams bordered by true drift terraces. They may be called backwater terraces. When in the Ohio, Muskingum, Hocking, etc., rivers, the water in the drift era stood eighty or ninety feet higher than at present, the backwater would set back up all the tributaries. In this still water the sand and sediment brought down these tributaries were deposited; or, in other words, the still-water areas were silted up, as mill ponds often are. When afterward the main streams fell to their present level these affluents cut through the backwater beds and carried away much of the soft materials, but left in many places fringing terraces which tell very plainly how they were formed. In these backwater terraces we find no true drift sand or gravel. The beds are entirely of home origin. Such terraces I have seen in the Little Scioto River, above its junction with the Ohio at Sciotoville, on Duck Creek, and on the Little Muskingum River in Washington County, and on Sunday Creek in Athens County. I have no doubt they are to be found on a large number of streams." (Geology of Ohio, ii, 1874, p. 444.) This record shows a delicacy of observation and a skill in physical interpretation that have impressed me as exceptional and admirable.
Taking up again the comparison of the two classes of constrained streams, it is seen that the channels of the first class were cut down to so gentle a grade by the filtered glacial streams that they are now not infrequently found to be filling up, and lakes are forming in them; but the valleys of the second class were filled so high by the gravels deposited from the surcharged glacial streams that they are now being terraced, and the lakes that were formed on their lateral tributaries are now discharged