[43] The Silt Problem and Storage Reservoirs. 1 1 
the percentages found from Professor Goss’s volumetric determinations 
at the end of one month by one-sixth, in order to have at least as large 
a percentage as would exist at the end of one year, we shall have for the 
sample collected on July 23, 1899, about 22 per cent, volumetrically, 
whereas the gravimetric method yielded only 10.78 per cent. For the 
sample collected August 13, 1899, the volumetric determination, reduced 
to the probable basis of a year’s settlement, yielded 28 per cent., while 
the determination by weight yielded only 12.4 per cent. With the third 
sample, collected December 4, 1900, the reduction is much greater, as 
would be expected from the smaller quantity of silt present. At the 
end of one year the percentage by volume would be a little more than 
3.7, whereas by weight the percentage was only 1.15 per cent., which is 
less than one-third of the volumes found for the silt gravimetrically. 
In a few determinations on the Brazos and Wichita river waters about 
the same relation was found, from which I have concluded that when the 
quantity of silt is determined by weight the results are only about one- 
third, or at most one-half of the quantity determined by volumes. 
In regard to the method of determining the sediment volumetrically, 
it may be objected that the sediment at the bottom of the laboratory 
tubes is under a very different pressure from that which would exist at 
the bottom of a deep reservoir. However, I hope to be able to test the 
effect of increased pressure experimentally during the coming year, but 
I do not expect it to make very much difference, for the silt at the 
bottom of the tubes appears to become as compact as that which 
remained in the lake at Austin after the failure of the dam. Two 
months and a half after the break I visited the lake and observed the 
silt still remaining. In some places a crust had formed on top hard 
enough to support a man’s weight, but everywhere the sediment was full 
of crevices, due to shrinkage in drying — many of these being several 
feet in depth. Below the crust the sediment was still wet and soft, and 
did not appear to have been very much compacted, notwithstanding the 
softer upper portions must have been swept out with the rush of water 
that followed the break in the dam, and which continued for a long 
time afterwards. 
From what has been said, it will be seen that in order to extend the 
life of a reservoir as much as possible great care in its location should 
be exercised. The reservoir should be located on a stream of small 
flow, or on one which carries little sediment, and the capacity of the 
reservoir should be so large that the inflow from a stream that carries 
as much as 1 per cent, of silt, determined . volumetrically, should be 
such that it would fill only once, or at most twice a year. Sluice gates 
at the bottom of the dam have not so far been found to prevent the 
deposition of silt, though if left open during the rising and flood stages 
of the river they may accomplish something in maintaining a current 
