SECTION OF IRRIGATION AND METEOROLOGY. 
13 
deduces an expression for the leakage through embankments, which 
seems to be proportional to the square root of the cube of the 
depth. Much of the loss may take place through the bottom, and 
this would tend to make the leakage increase at a higher rate. 
Through the courtesy of John C. Trautwine, Jr., Chief of the 
Bureau of Water, Philadelphia, I am enabled to quote some recent 
and valuable observations on one of the reservoirs belonging to that 
city, and which clearly show the effect of increased depth. The 
Queen Lane reservoir has considerable trouble on account of leak- 
age, and during the summer of 1897 had been lined with asphalt, 
the lining being completed August 16th. 
The diagram shows the observations at different depths, the 
two curves representing the losses before and after partially silting 
the reservoir by pumping into the reservoir water laden with an- 
thracite coal dust. 
November, 1897. 
March, 1898. 
(Before Silting.) 
(After Silting.) 
Depth . 
Loss per Day. 
Depth. 
Loss per Day . 
15 feet. 
.29 inches. 
20 feet. 
.15 inches. 
20 feet. 
.54 inches. 
25 feet. 
.24 inches. 
28 feet. 
. 32 inches . 
j 30 feet. 
.40 inches. 
These observations show in both cases a more rapid increase in 
the leakage with depth than would seem to be indicated by the un- 
certain theory. 
That a portion of the rapid increase is due to some change in 
the conditions is shown by still more recent observations. Under 
date of April 7th, Mr. Trautwine adds that observations, extending 
over ten days, with a nearly constant head of about 30 feet, have 
shown an increase in the daily loss to .55 inch. 
Since that date the reservoir has been drawn down to 20 feet, 
and under date of April 16th, Mr. Trautwine writes that the aver- 
age daily loss at that depth for eleven days was .28 inch, or about 
double the amount shown in March. 
These interesting facts would seem to show that the conditions 
remaining the same, the effect of an increase in leakage due to an 
increase in depth, would be at a much less rapid rate than indi- 
cated in the curve shown. 
As an explanation of the increase in rate with increased depth, 
and the continued increase after the depth is reduced, Mr. Traut- 
wine suggests that there may be a velocity of percolation at which 
the sediment-bearing water ceases to deposit its sediment in the 
pores and begins to carry away that which has already been de- 
posited, thus permitting an increase under large heads with time. 
