376 Chamberlin and Salisbury — Relationship of 



loess and gravel have been carried down slopes and deposited 

 together at their bases. But none of the gravel beds of this 

 category belong to the Orange Sand formation, or to the 

 formation of bluff gravels of Safford, and their existence has 

 no bearing upon the conclusion cited above. 



It remains to indicate briefly the relationship of the second 

 glacial deposits to those of the first, in the region under con- 

 sideration. Starting from the points where the edge of the 

 ice sheet crossed them, trains of gravels stretch down the valleys 

 which served as avenues of discharge for the glacial waters. 

 From the terminal moraines of the second epoch in which 

 they head, these gravel trains, in the form of valley plains, 

 extend down the valleys for great distances. Into these plains, 

 post-glacial streams have sunk themselves, and such portions 

 of the plains as have escaped removal constitute the gravel 

 terraces which are so generally found in the valleys heading 

 within the area of second glacial drift, from the Atlantic to 

 Dakota. From the position of these terraces in relation to the 

 terminal moraines, there can be no doubt as to their second 

 glacial age. Their constitution too, in comparison with that 

 of the material of the same grade of coarseness in the adjacent 

 second glacial till, confirms the conclusion to which their posi- 

 tion leads. Traced down stream, these terraces decline, some- 

 what rapidly at first, but soon more slowly. There is also a 

 change in their physical constitution as followed down the 

 valleys. The coarse gravel issuing from the moraines soon 

 gives place to finer, and this ultimately to sand. Often there 

 is but a single terrace, representing the level of the glacial 

 flood plain, but often also, there are terraces at lower levels, 

 representing subsequent stages of post-glacial river degrada- 

 tion. 



These terraces of the Mississippi Valley are of especial 

 significance in this connection, since this valley traverses the 

 region under consideration. !Not only does the altitude of the 

 main terrace above the sea decline southward from the moraine, 

 but its elevation above the stream likewise diminishes in the 

 same direction, until it disappears altogether below the pres- 

 ent flood plain of the Mississippi.* We have not seen the 

 terrace south of the limit of glacial drift in Illinois, though 

 terrace remnants near the mouths of tributaries occur at about 

 that latitude, and may exist still further down the valley. 



In the Ohio Valley at Louisville, there is a corresponding 

 terrace, built up of material brought down by tributary streams 

 heading within the area covered by second glacial ice. This 



* Below the point where second glacial terraces have been identified, there are, 

 in places, low plains slightly above the present flood plain of the Mississippi, 

 which possibly are to be correlated with these terraces. 



