Ii6 
Journal of Agricultural Research 
Vol. IX, No. 4 
The action of this device depends upon an adaptation or extension of 
Venturi’s principle to the flow of a liquid in an open channel. As water 
passes through the flume there is a slight surface slope in the converging 
section, a rather sudden depression in the “throat” section, and a rise 
in the diverging section. The actual loss of head is small. The deter¬ 
mination of the flow depends upon the velocity and wetted cross-sectional 
area at two points in the flume, and two gage readings, therefore, are 
necessary. One gage has been arbitrarily located upstream from the 
throat a distance equal to two-thirds the length of the converging sec¬ 
tion, to avoid possible influence due to contraction currents nearer the 
entrance to the flume; and the other gage has been located at the middle 
of the throat section, in order to obtain the greatest possible difference 
in elevation of water surface. The zero of these gages must be at the 
elevation of the floor of the flume, and it is especially important that 
the zero of the gages be at exactly the same elevation. The difference 
in heads, H dt is a more important factor in determining the discharge 
than the depth of water in the channel, H a or H 5. 
Still boxes, or gage wells, are necessary for accurate readings of the 
water levels, because of the comparatively high velocity of the water 
flowing through the structure. Field tests on small Venturi flumes 1 
indicated that readings taken to the nearest 0.01 foot on staff gages placed 
at the proper locations inside the flume, with the face of the gages coun¬ 
tersunk flush with the surface of the side of the flume, would give an 
accuracy of measurement sufficient for general purposes. This would 
overcome the necessity for using gage wells, but recent tests made in 
the laboratory show that such staff-gage readings do not agree with 
readings taken in the gage wells when there is enough fall in the carrying 
channel to give a high velocity of flow through the flume, in which case 
H d is a considerable amount. Until more is known of the accuracy 
of gages under different arrangements, caution should be used. 
Instrument makers are at work on an automatic register to make 
graphs of the water elevations at the two gages, both records to appear 
on a single sheet. An integrating register would be most desirable, but 
the complexity of the law of flow through the flume certainly would 
require a complicated instrument. 
The effect of the velocity of approach is automatically cared for in the 
device, and the formula takes account of the velocity of the water at 
each gage. The experiments indicate that the Venturi flume will be free 
from interference due to changes in the .canal section, such as occur often 
from sand or silt accumulations or aquatic growths. Such obstructions 
make the use of the ordinary rating flume very troublesome, if not quite 
impossible, but these obstructions result only in changing the relative 
gage readings of the Venturi flume without altering the calibration of 
1 Tests made on the North Platte Project, United States Reclamation Service, Mitchell, Nebraska, under 
the general direction of Mr. Andrew Weiss, Project Manager. 
