Nov., 1907.] Aggradation and Degradation of Valleys. 
*93 
valley than in the lower. In the latter the stream spread out 
lazily in a shallow sheet covering nearly the entire width of the 
valley floor. Here little if any erosion was taking place, because 
the stream was essentially at base level. Sediments picked up 
above were transported across this area without modifying the 
lower valley, and into the lake to extend the delta. As often 
as the delta was built out to the plug, it was artificially removed. 
In order to have the lower part of the valley degraded, it was 
found necessary gradually to lower the plug in the lake, just as, 
reversing conditions, it was necessary while aggrading the valley 
to increase its height. This done, terraces were formed by the 
stream’s meandering in the lower course, as well as in the upper. 
The rock spurs mentioned at first, jutting out into the valley, 
resembling in perspective the entering spurs from alternate sides 
of a valley as seen in mature, dissected plateaus, were built to 
test the theory of “defended cusps” as set forth in a paper by 
Davis. 2 Our results supported his theory perfectly, for while 
the stream swung back and forth across its meander belt, leaving 
terraces at the limits of each meander, these terraces were very 
often destroyed by the next migration until the cutting had 
reached sufficient depth to discover the rock spurs, when subse¬ 
quently all terraces, higher on the bluffs than the spur encoun¬ 
tered, were preserved. Also in many cases terraces down 
stream from the spurs were sufficiently defended for preservation. 
In Fig. 9, A, B, and C represent terrace tops, t and m, the 
ledges encountered by the stream, and n, a continuation of the 
effective spur seen above. While the stream was flowing on 
terrace C the ledge t protected terraces B and A from being 
destroyed. 
The experiment showed also that the succession of terraces 
with narrowing strath on each lower level was not due to a 
decrease in the volume of the stream, for we were careful to 
maintain a constant stream; and, while there were temporary 
variations, it is probable that every hour’s work was done by 
about the same quantity of water, and certain that no appre¬ 
ciable decrease in the volume occurred during the degradation 
process. The terraces were made by the persistent systematic 
swinging of the stream back and forth across its flood plain. 
Where the stream encountered the ledges it was restrained in its 
lateral cutting and each lower level presented a narrower valley 
than the one above, but where no ledges were met the stream 
in the end usually undercut the older terraces and destroyed 
them. The stream in this erosion stage of its work, however, 
did not change its course nearly so often as it did during the 
stage of sedimentation suggesting, as has been shown by Griggs, 3 
2 Bull. Mus. Comp. Zool. Harvard. 1902. Geol. Series Vol. V. 
3 Bull. Amer. Gcog. Soc. 1906. Vol. XXXVIII. pp. 168-177. 
