THE FLOW OF WATER IX DRAIX TILE. 15 
ing the lieight of the water surface outside of the tile at the joint. 
In some cases the water level outside the tile would remain steadily 
0.1 or 0.2 foot higher on one side of the tile than on the other, and the 
surface inside the tile would be very turbuJent and wouJd seem to 
bear no relation to the elevation of the water surface outside the tile 
joint. Such phenomena were most conspicuous when the depth of 
flow was between half and full depth, and with the high velocities due 
to the steeper slopes. The phenomena seemed to depend upon the 
presence of air in the tile, as thev disappeared largely when the tile 
were completely filled, so that all air was excluded. 
METHODS OF CONDUCTING TESTS. 
A test was always begun at the least depth of flow. Six men were 
needed to conduct a complete experiment at one grade, which required 
from 3 to 6 hours, depencUng upon the number of depths of flow 
tested. One man cared for the pump and engine, one read the upper 
hook gage, a third was stationed midway the length of the fliune at 
a piezometer tube, another was stationed at the outlet to adjust the 
height of the movable dam, and a fifth man read the lower hook gage. 
The engineer in charge usually operated the valve controlling the 
supply of water to the upper weir tank, and watched the upper 
piezometer tubes. 
The engineer announced the depth of flow he desired to obtain to 
the man stationed at the dam. The gate valve in the supply pipe 
was partly opened and the piezometer readings noted. The dam was 
then raised or lov/ered to secure the correct depth of flow at the 
piezometer tube near the outlet, special care being taken not to get a 
greater depth than desired there. The observer at the upper hook 
gage called out the various gage heights at short intervals, that the 
water supply might be regulated properly, and when the desired depth 
in the tile was obtained, sufficient time was aUowed to determine that 
the depth over the weir was constant. The observations at the 
upper, middle, and lower piezometers indicated when the flow was 
steady throughout the tile Ime. Wfien the flow was steady at the 
proper depth, the signal was given and each of the two hook-gage 
readers made record of the readings at his station every 30 seconds. 
Meanwhile, the engineer in charge passed along the flume, recording 
the readings of aU piezometers in succession; the observer at the 
lower end of the flume went to the upper end and then recorded the 
piezometer readings in order, following just 2 minutes behind the 
engineer's readings ; the observer at the middle of the flume watched 
the piezometer there to report if any considerable fluctuation indi- 
cated that the test should be run again. If the depth over the upper 
weir remained constant throughout the test, the engineer proceeded 
to obtain the next depth of flow; if the weir readings varied, the test 
