150 The N.Z. Journal of Science and Technology. [May 
which have been filled either with alluvium or with glacial drift. 
A river re-excavating a valley of this kind and taking a course across 
a buried spur has its downcutting so retarded thereby that it develops 
a graded, fully mature, and broadly opened valley farther up-stream. 
Such a river, in its graded reaches, continues during downcutting to 
follow a meandering course over a flood-plain ; the meanders migrate 
down - stream, and the meander belt swings from side to side of the 
valley-floor. Since, however, the stream is degrading, each time the 
meander belt approaches the valley - side its floor is at a lower level 
If it quite reaches the valley - side 
it completely cuts away the former 
flood - plain, but if not it leaves 
a remnant as a terrace. Fig. 10 
illustrates this method of terrace- 
formation. Block A represents the 
river flowing on a broad valley-plain 
in an alluvium - filled valley. In 
block B degradation has begun, and 
a strip of the width of the meander 
belt has been incised to some depth, 
leaving a terrace of the same height 
on each side of the valley. In block C 
the meander belt, now more deeply 
incised, has swung towards the right, 
leaving as a second terrace a portion 
of what was its floor in block B. 
Block D represents a later stage, in 
which a third terrace has been left. 
It will be noted that the fronts of 
terraces developed in this way are 
concave, as they mark the farthest 
sideward migration of a convex curve 
of the meander belt. 
Such terraces as these are not com¬ 
monly found in the valleys of rivers 
that are degrading in homogeneous 
material, even though the material be 
soft, for in this case there is nothing 
to check the lateral swinging of the 
meander belt at the lower, levels, and 
so the higher floors are liable to be 
entirely cut away. When, however, there are bars of solid rock such as 
are provided by buried spurs in the soft alluvium of a filled valley, these, 
when they become exposed, prevent further swinging and so protect series 
of terraces both up-stream and down-stream from the rock ledges. Upon 
the ledges the concave fronts of up-stream and down-stream portions of 
terraces meet in cusps.* 
* The part played by rock ledges in defending terraces was first remarked upon 
by H. Miller (River Terracing, its Methods and their Results, Proc. Roy. Phys. Soc. 
Edin., vol. 7, pp. 263-305, 1883). It was further discussed by W. M. Davis, particularly 
with reference to the terrace patterns developed (River Terraces in New England, Bull. 
Mus. Comp. Zool. Harv., vol. 38, pp. 281-346, 1902 ; reprinted in Geographical Essays , 
Boston, 1909, pp. 514-86). 
than that of the previous time. 
Fig. 10.—Diagram to explain the de¬ 
velopment of terraces by a river 
degrading slowly in soft material. 
