480 Journal of the Asiatic Society of Bengal. [September, 1908. 
other words in want of tractability, as the lower Ganges is at the 
present moment; the slope of the bed of the Kosi River, measured 
which the Kosi mp nays its silt is probably on an average at least 
20 miles wide; if we accept.a strip of 5 miles on each side as the 
area on which. depositidn will occur, we must err on the right side 
in our calculation; in other words, we assume that the area to be 
raised is 100 miles i in length, and 10 miles in breadth. If from the 
ve figures we calculate the time which must elapse for the 
slope of the 100 miles of the Kosi under discussion to average 6 inches 
per mile, we shall find that abont 1,000 years is the answer to the 
calculation, I admit the fizures I have acc ase are not based on 
the resuts of observations, but, notwithstanding this, they give not 
un-interesting results, 
he period is, in all probability, much shorter than that which 
will actually onde before the Kosi ewe as sab a S sin‘. as 
the lowe: ges is in to-da ‘the plain 
in the mateo and th uy 5 Seageemniee se anne 
cted; if the depression equa: building power id 
the river, matters will remain stationary Rat one side. Divoselis 
eee Litag the other. From a human point of view, therefore, if 
we me that the plains are sti gy paali our estimate may be 
aapled.: Sah avenler eis ener omy nfinity. From the above, it 
will, I think, be: Gime has: not yet come for the 
rigid training of the Kosi. River. 
ore any practical attempt can be made to save “those 
in Purnea and Bhagalpur from. the ravages of the Kosi, a 
0 cross sections of the river - surveyed normally’ to ‘the 
current, one ae a point as close to the exit from the Siwaliks as 
possible, and another at the railway bridge; the latter may be 
obtainable from the B. & N. W. Ry. authorities, If possible, a third 
section should be measured midway between the two just men- 
tioned. 
2) At each section, measurements of (a) wedi bed_, 
(b) rate of current; (c) amonnt of silt carried; (d) che mical 
examination of silt carried at different phases of flood level. of the 
stream, should be eee oe at least one year. (2) Should a 
taken monthly; (6 (c) daily at a given time; and | 
all differences of height OF stream va intervals of 6-inch change. 
(3) From data olde described, the mean annual disc 
of the river (a) in water, (6) in silt, should be computed. 
4) The rate at while sh the plains are being depressed and the 
Siwalike raised, should be found experimentally ; this will. entail 
the discovery of several points which lie between the Siwaliks and 
their complement, the plains which do not alter in altitade. a 
several lines of really careful levels were run in such a way as 
conyerge from masonry points on the hills on sha act cee ad the 
NAR ip serine nie 8 
