Measurement and Division oe Water. 33 
of the sill is not sharp or is slightly rounded on the upstream side, 
there is an appreciable effect. A serious fault of the weir is, that 
carelessness in making or neglect in maintaining it, tend inevitably 
to increase the discharge. It is not to the interest of the consumer 
to see that the proper conditions are maintained. 
The measurement of the depth of the water is difficult to make 
accurately. Frequently this is done by placing a post some distance 
above the weir and the depths measured on it. This should be at 
least six feet above the weir. As the water is constantly subject to 
fluctuations this can not usually be done with sufficient accuracy. 
The best place to measure the water would be in a well to one side 
of the basin and connected with it by means of a small tube. This 
well may be nearer than six feet to the plane of the weir because 
the pressure of the water will bring the water level with the surface 
further up stream. It should be from 3 to 12 inches in diam¬ 
eter or square, depending on how it is to be used. A float may be 
made with an arm extending above the surface graduated to read 
zero when the water is level with the crest of the weir. The rules 
commonly to be had in the market are for the use of mechanics and 
divided to i6ths of an inch. Rules divided to decimals of a foot 
would be better for this purpose. 
These conditions all pertain to the rectangular weir, as well 
as to the trapezoidal. The Cippoletti or trapezodial form of weir 
is simply different in the form of the opening. It has been shown 
that the side contraction is greater as the depth of the water over 
the weir is greater. The broadening of the opening is such as 
to automatically allow for that contraction, so that the calculation 
is based upon the length of the weir without allowing for con¬ 
traction. 
SUBMERGED OR DROWNED WEIRS. 
Many times the water backs to or over the crest of the weir. 
This is sometimes accidental, and sometimes is due to the fact that 
it is necessary to run water on land close to the main canal without 
sufficient drop. While it is not desirable to depend on submerged 
weirs, as these are called, yet within moderate limits the results 
of such weirs can be depended on and the ordinary tables used. 
The first effect of backing over the crest is to increase the 
discharge, because the air underneath the falling sheet is displaced 
and the air pressure tends to increase the amount more than the 
decrease due to the discharge under water. Then when the sub¬ 
mergence is about one-seventh of the total head, these effects balance 
and the discharge is again the same as that given by the tables. 
With greater submergence there is more decrease, so that when 
the submergence amounts to one-quarter of the head the discharge 
