THE FLOW OF YVATEE IX' DEAIX TILE. 7 
To decrease the amount of work necessary to adjust the grade of 
this continuous channel, the flume was rotated about its longitudinal 
center. Thus, when changing grade, one half of the flume would be 
lowered while the other haK would be raised. The flume could be 
set to any grade up to 1.50 feet in 100 feet. 
To enable the workmen to determine whether the flume was at the 
proper grade, graduated wooden strips (A, PL VI, ^g. 1) 2 inches 
wide, 0.5 inch thick, and several feet long were placed on each side 
of the flume at each bent. The difference of elevation between 
various grades at each bent had been previously computed, and these 
differences were marked on the gage strips with the corresponding 
grade number. Thus, when the proper mark apjDeared at the cross 
board thi^ough which the gage strip ran, the workmen knew that part 
of the flume to be at the desired grade. At points where the required 
change of elevation was considerable, the flume was not raised or 
lowered the entire am.ount at one time, but was changed by successive 
increments of only a few inches. Thus the amount of stress on the 
flume was lessened, and the liability of leakage through the possible 
springing of the planks was eliminated. 
The grade of the flume was checked with an engineer's level im- 
mediately before each experiment, to eliminate all possible errors 
from inaccurate adjustment or from settlement of the vertical posts. 
LAYING THE TILE. 
The tile were laid on earth in the flume as in actual practice. 
This earth in the bottom of the flume was about 7 inches deep. It 
was placed in layers, 2 inches at a time, and each layer was thoroughly 
tamped so that the bed on which the tile rested would not settle. 
At fii^t a line was stretched along the flume immediately over its 
center and about 3 feet above the grade, and this line was used to 
grade the bed for the tile. It was soon found, however, that the gage 
line was in the way of the workmen, and another method for grading 
the tile bed was adopted. The material for this method consisted of 
a 30-inch strip, 2 inches wide and 0.75 inch thick, and a gage stick of 
the same size but 17 J inches long. The workman laid the strip 
across the top of the flume and, holding the top of this gage stick 
flush with the top of the cross arm, determined whether the invert of 
each tile was at grade (P. VI, flg. 2). 
COVERING THE TILE. 
While blinding the tile, an engineer was constantly in the flume to 
oversee the work and prevent any tile from being pushed out of Hne. 
Fine earth, free from large clods, was used for blinding^ the inspector 
tamping the earth on each side of the tile with his feet. Thus any 
appreciable movement or current of water through the earth on 
