i8 
LOSSES FROM CxVNALS BY SEEPAGE. 
Mr. J. Keelhoff made some experiments on the absorp- 
tion of small ditches.* 
From the facts given by him, we find that the loss in 
the sandy soil of the Campine from irrigation ditches, lo 
inches wide with water ii inches deep, was over lo feet in 
depth in 24 hours; but when the depth of water in the ditch 
was but two i ches, the loss was reduced to six feet per day. 
In the distributing laterals, 10 inches deep and two feet 
wide, the loss was over four feet in depth per day. In the 
principal lateral, with water 2 feet deep and 8 feet wide, the 
loss was over 2 feet per day. One reason for less loss in 
the last case, though the water was deeper, is that the bot- 
tom remains undisturbed from year to year. At the time 
of the test the silt had not been removed for four years. 
I'he other ditches were cleaned annually, thus giving a 
raw surface for the water to pass through. 
Geo. VV. Rafter, C. E. in a report on the water supply 
of the W stern division of the Erie canal, f refers to a 
number of determinations of the losses from seepage and 
evaporation on stretches of that canal. 
On a section of 18 miles near Schenectady through an 
alluvial soil containing a large proportion of vegetable matter, 
and leaky in places, the loss as mea'^ured by J. B. Jervis in 
1824, was 2 cu. ft. per second per mile. The canal was 28 
feet wide on bottom, 40 feet wide on top and 4 feet deep. 
Idiis is equivalent to a loss in depth of 10 inches in 24 hours 
over the whole surface. 
Mr. David S. Bates in 1 823 concluded that a mile of new 
canal, such as the Erie then was near Brockport, would re- 
quire cu. ft. per sec. per mile. This included evapora- 
tion. I'he dimensions of the canal are presumably the same 
as the above, in which case the loss would be equivalent to 
8 inches in depth per day. On the Chenango canal in Aug. 
1839, the amount was found to be i.oq second feet per mile, 
corresponding to a depth of 6 inches in 24 hours. 
On the Erie canal near Wayneport, in 1841, in a dis- 
tance of 8 miles, when the soil was open and porous, the 
loss was 1.8 cu. ft. per second per mile; on the Clyde level, 
a length of 28 miles, with more retentive soil, the loss per 
mile was only .6 cubic feet per second. These correspond 
to depths of 9 inches and 3 inches per day respectively. 
In comparing with the results found on irrigation ca- 
nals, it should be remembered that the conditions on the 
*Traite Pratique de rirrigation des Prairies. 2d ed. 
t Rpport of the State Engineer and Surveyor of N. Y., 1890. 
