8 
Colorado Experiment Station 
with the surface, do not give the best results. It was first thought 
that two points of observing the heads would be sufficient to 
definitely establish the conditions of flow for rectangular flumes, 
but due to the switching of the current in the diverging section, it 
was found necessary to observe the head at identical points on each 
side of the flume. The introduction of additional points of observ- 
ing the head is not a desirable feature, but is necessary to insure the 
dependability and accuracy of the device. Table 1 gives the errors 
introduced into the discharge by determining the heads from one 
set of gage wells which register the depths on one side of the flume 
only. To avoid the possibility of uncertainty in the readings and 
to obtain a more correct determination, it is suggested that the 
arrangement of wells and connections be made as shown in Figure 1. 
It has been suggested that a narrower diverging section might 
so modify the downstream conditions as to permit of reliable results 
from reading one side only; however, this restriction would no 
doubt greatly increase the mean velocity at the exit of the flume, 
thus causing a scouring action in the channel immediately below the 
structure which would require the installation of protection against 
this action. 
The velocity of approach to the Venturi flume has no effect 
upon the accuracy of this device. When a certain quantity is flow- 
ing through the flume, drawing its supply from a still water basin, 
there will be. a definite head at the gage connection in the converg- 
ing section and in the throat, also a definite difference in head. If 
this same discharge were to approach the flume with a velocity, the 
upper gage reading* would be less as would also the throat reading, 
and the difference in head would be greater. Since the upper head is 
less, the difference must be greater tO' permit of the same discharge, 
and because of the increased velocity, the head at the throat section 
is automatically decreased to balance the change in the velocity of 
approach. 
In the Venturi flume the loss of head, due to the structure, is 
small except when the difference in head is great. Under normal 
conditions met with in practice, the velocity head is recovered when 
^the water leaves the lower edge of the flume. However, when the 
difference in head is large, the water will maintain its high velocity 
through the flume and will not become normal until some obstrpc- 
tion, as an irregularity in the downstream channel, causes the for- 
mation of standing waves and the change from a high to a low 
velocity. 
