30 
MEASUREMENT AND DIVISION OF WATER. 
the discharge; Cippoletti therefore increases Francis’ co¬ 
efficient by 1 per cent, to allow for the preponderance of 
accidental causes increasing the discharge, thus making 
the equation become 
Q = 3.36§ L FI 2 
where Q, L, H, represent the same quantities as on page 21 r 
except that L is here the actual length of the weir. 
This form seems to possess some very advantageous 
qualities, rendering it much more simple in use than the rec¬ 
tangular weir and dispensing with some of the objections 
which have not rendered the weir popular. The effective 
length to be used in calculation is the same for all deptlis r 
while in the rectangular weir, with complete con¬ 
traction, every change in depth alters the effective 
length and makes the calculation inconvenient and 
laborious. The discharge also being in proportion to the 
length, very nearly, makes it convenient for application 
with such a device as the spill-box, where the opening 
may be made of this form instead of that used by Graves. 
Then, if one person is entitled to one-half as much as an¬ 
other, the sill may be made half as long, or a moveable 
gate may reduce the width with an error probably very 
small, and if any error it is in favor of the smaller user. 
The amount ot water available and our apparatus at 
present available for testing the flow of water is not suf¬ 
ficient to test this to any great extent. The flow over 
6-inch and 12-incli weirs has been measured. These seem 
to indicate that the correction to Francis’ formula for these 
lengths would be nearer 2 than 1 per cent. It would be 
less for larger weirs. The experiments are sufficient to show 
that the Cippoletti weir may be used under the conditions 
stated, with confidence that the error will not be likely to 
exceed 1 per cent. 
