10 
Colorado Experiment Station 
The Venturi flume is gaining in favor as a practical device, and 
the fact that it possesses many desirable features and few objec- 
tionable points bids fair to establish it as one of the standard 
measuring devices for practical irrigation purposes. The present 
design is not free from all objection, and is not claimed to be abso- 
lute in its operation, but when compared with other means of meas- 
urement it is found to be within reasonable limits of accuracy. 
RECTANGULAR VENTURI FLUME 
Since the publication of the first report on the Venturi flume, 
there has been a great number of these flumes built and installed for 
use in the measurement of water to irrigators. So' far the greater 
per cent has been the rectangular type, and it appears to be some- 
what more popular than the trapezoidal flume; however, there has 
been a number of the trapezoidal Venturi flumes installed, the cali- 
bration having been made by the interested parties. 
The preliminary investigations of the Venturi flume resulted 
in a selection of certain ratios and dimensions applicable to a stand- 
ard plan for the rectangular type. The assumptions as to dimen- 
sions were based on conclusions made evident by the behavior of 
model flumes of various ratios, and for uniformity the length of 
throat for all the sizes tested has been made one foot. The angle of 
convergence and divergence is equal and constant for the various 
sized flumes calibrated. The dimensions for the rectangular Ven- 
turi flume are shown in Figure 1. The calibration of this flume 
has covered a great many individual tests and includes the 1-, 1 %-, 
2-, 3-, and 5-foot sizes. The measured discharge has ranged from 
less than 1 to nearly 400 second feet in the calibration of this type 
of flume. 
The diagrams and tables pertaining to the rectangular flume 
contained in this bulletin are based on the data collected at Cornell, 
our original data, and the more recent calibrations made at our 
field laboratory at Bellvue, Colorado. 
As previously stated, our recent investigations indicate the 
desirability of providing the flume with four gage wells, two on 
each side of the flume. Tests were made to determine the per cent 
of error resulting from taking the readings in only one set of gage 
wells. Table 1 gives the results of these tests. In this table, the 
deviation from the true amount is either plus or minus, depending 
upon which side of the flume the observation wells are located ; or 
in other words, which side the current is flowing in the diverging 
