JAGGAR: EROSION AND SEDIMENTATION. 301 
mental lines of investigation connected with land drainage. ‘These 
are suggestive and qualitative, as the experiments have not yet 
progressed to the quantitative stage. 
Drainage areas and stream shadow have been discussed on p. 289; 
the presence of drainage surfaces essential to arborescence on p. 289; 
the importance of volume as contrasted with slope on p. 291; the in- 
verse relation of deposits to beds eroded on p. 291; parallelism and 
rhythm on p. 295 and thereafter; the comparison of capture and shadow 
on p. 296; and in several places reference has been made to the under- 
mining of surfaces whereby tributaries arise and whereby a propaga- 
tion both upslope and downslope of certain rhythms may start from an 
intermediate region. In the trickle-pattern on glass, V-shaped deposits 
pointing upstream form in a mouthward (p. 292) zone of overload. 
In most of the experiments there is an upper zone of water accumula- 
tion without channelling. Parallelism and rhythm in distribution of 
tributaries and spurs may by the mechanism of arborescent drainage 
be satisfactorily accounted for, without any influence of parallel 
rock joints (p. 297). 
In Plate 5, figs. 1 and 2, and Plate 6, a marked characteristic is the 
presence of much larger portions of the original surface uneroded 
near the lower edge of the models than in the middle right-and-left 
zone of many tributaries. ‘The general surface is more lowered in 
the middle zone than in the upper or lower zones. The middle 
zone is the region of maximum corrasion, maximum removal of 
material, and maximum maturity of topography. The interstream 
divides have been lowered below the. original surface level, whereas 
the original surface level is still preserved in the flat interstream up- 
lands of the lower part of the models. Hence the general upland 
surface is bevelled headward for a certain distance from the region of 
the mouths of the streams. This would give it a flat catenary profile 
from the headward district mouthward. 
This profile would be uniform wherever measured. In Plate 5 nine- 
teen or twenty stream depressions would be crossed from right to left 
in the middle zone and about twelve in the lower zone. — Interstream 
divides in each case rise to a common level. The relief is greater in 
the lower zone. This is contrary to the prevalent conception, that 
would expect least relief and greatest maturity in the lower zone 
(Tarr, 1898). Bevelling and uniformity of crests in the same zone 
have been discussed by Tangier Smith (1899). He has developed a 
law of slopes as follows: summits follow slopes, and slopes are depen- 
dent on the rate of cutting of the streams at their foot. “Tf the alti- 
