14 BULLETIN 85J:, U. S. DEPAETMEXT OF AGRICULTURE. 
ACTUAL DEPTH OF FLOW. 
Tiie depths of flow in the tile lines tested were measured by means 
of the 12 piezometer tubes distributed along the flume, as previously 
described. 
As first laid, the tile at the lower end of the experimental line dis- 
charged into the open air \\'ith a free drop of several inches. This 
produced a backwater curve of the drop-off type which extended 
back for a considerable distance into the tile line, decreasmg the 
depth of the water near the lower end. For the steeper slopes this 
effect was much extended, and, indeed, in extreme cases reached 
throughout the length of the experimental line: Such a condition 
was objectionable for two reasons: fii^t, the hydraulic gradient under 
such ckcumstances would be represented by the slope of the water 
surface, which was then somewhat greater than the grade of the tile; 
second, since the condition was not one of uniform flow, it would 
become necessary to take account of the change of velocity at differ- 
ent points in the tile, with the corresponding changes in Yelocity head, 
in determining the head consumed in overcoming friction. Smce these 
additional compHcations were unnecessary and objectionable, the 
drop-off curve was eliminated by installing a low. movable dam 
(shown in PI. Ylll, fig. 1) just below the lower end of the tile line. 
By adjusting the height of this dam; the water surface at the outlet 
could be maintained in close agreement with any desired depth 
throughout the experimental line. 
The water entered the upper end of the tile through a conical 
entrance pipe (PI. Ill, figs. 1 and 2) designed to give an entrance 
velocity approximating that of the steady, uniform flow in the tile 
line. But it was found impracticable to adjust the entrance velocity 
exactly to that of the line, with the result that the upper 50 feet 
of tile were required to bring the velocity' to the condition of uniform 
flow, and the piezometers at the upper end would not always agree 
with the others along the tile. With this exception, the readings of 
depth in the various piezometers along the tile line could generally 
be brought into satisfactory agreement. 
With the tile only partly fuU, there were occasional quite erratic 
readings on some piezometei's. These indicated imusual disturb- 
ances witliin the tile line. When through the warped or elliptical 
shape of the tfle the joints do not fit closely, a portion of one tile 
at the joint may project inwai'd in such a way as to present a square 
obstruction against the edge of the moving stream of water. Violent 
impact of the water against such an obstruction produces a marked 
disturbance of the stream, and is indicated by extensive ripples and 
foam on the water surface which ma}^ pei^sist for several feet down- 
stream. Several such cases were carefuUy examined by uncovering 
the tile and inspecting the water surface within, ^s well as by measur- 
