Dodge.J 264 [March 21, 
portions of tlie westoni part (jf llic Lriiited Stiit-.'s, tlicii the 
stroums ol the region would have a slightly decreased volume on 
account of dessicatioii and being furnished with a co Distantly 
decreasing load from the valley sides, would find their carrying 
jiower still ahead of the work supplied to them. Therefore they 
would begin to incise their channels in the alluvial tilling dropped 
dining a j)revious time of greater precipitation. Terraces 
might thus be formed when the rivers deserted their old flood 
])lain levels. Such terraces have been described by Drew^ from 
the rivers of the Himalayas. It may seem at iirst that it is not 
legitimate to consider that climatic changes of such magnitude 
would take place during one cycle of a river's growth ; yet the 
knowledge that Lake Bonneville has been several times filled and 
emptied during the present incomplete geographical cycle shows 
us that such climatic oscillations are perfectly natural in the history 
of a river. The infrequency of such occurrences, however, makes 
it advisable to consider terraces due to this cause subnormal 
rather than normal. 
A second way in which a subnormal terrace might be formed 
by a river is as follows: suppose that a growing talus slope, 
alluvial cone, or a lanti slide, all of which are perfectly normal 
features of a river's development, should build a barrier across 
a stream, damming it back into a lake. Such a barrier would 
act as an important control on the work of the river as tlie 
temporary base-level of the country up stream would be the 
level of the barrier as long as that barrier existed. Conse- 
quently the stream would lose much of its carrying power as 
it approached the level of the barrier and would deposit much 
of its load in the temporary lake formed back of the dam. 
When in the com'se of time the river should succeed in cutting 
down through the barrier, it would again acquire a greater gradi- 
ent and would thus be able to carry a greater load with a given 
volume. It would therefore cease to deposit detritus above the 
barrier and would begin to cut its channel down into the alluvial 
deposits already dropped there, and in so doing might possibly 
leave terraces. 
Furthermore the formation of a barrier might have an impor- 
tant effect on the down stream portion of the river, for it might 
' Alluvial and laciislriiir defxisits aiul i;hicial records of the iqiiier Indus basin, 
i.tiiarl. joiirn. grul. soc. l/md., v. -Jt'.), p. 141-471, 1873. 
