14 BULLETIN 852, U. S. DEPARTMENT OF AGRICULTURE. 
voir X blown up with the bicycle pump until under a pressure much 
in excess of that existing in the pipe under test. At a definite noted 
time the color is injected into the pipe through the cock V. The 
gauge readings were taken simultaneously with the passage of the 
color through the pipe. 
Color siphon. — Where the pipe discharged into a sand or division 
box (PL III, fig. 1), the first appearance of the color could be noted 
easily at the pipe outlet, even though 6 or 8 feet below the surface, if 
the water was clear, but it was impossible to determine when the last 
sight of color reached the end of the pipe, as the pool would then be 
filled with colored water. To overcome this difficulty a length of 
garden hose was calibrated as a true siphon. The inlet end of this 
siphon was then held in the pipe outlet and the last appearance of 
color was observed at the outlet of the hose. From the calibration 
it was found that an effective head of 1 foot required four seconds to 
discharge a volume of water equal to the volume of the interior of the 
hose siphon. Thus, if the outlet of this siphon was always just 1 foot 
lower than the water in the outlet chamber, then the last appearance 
of the color in the concrete pipe was just four seconds before its last 
appearance at the hose outlet. 
Air trap. — Since all water flowing in pipes contains more or less 
air in solution and sometimes even free bubbles of air, an accumu- 
lation of air in the piezometer tubes is to be looked for. In order to 
insure a water column free from air between the pipe and the gauge 
glass, the attachment shown in Plate V, figure 1, was developed. A 
solid water column follows the tubing aaa to the gauge glass, while 
air bubbles, taking the highest path, pass into bb. Any bubbles 
that might escape the first glass Y at b are given another oppor- 
tunity to escape at c. The air chamber d is provided with an air 
valve. In blowing out the pressure tubing , the valve V is opened 
wide to create a draft. The air valve at the gauge glass is opened 
very slightly. A relatively large amount of water escapes at V and 
a small amount at the gauge glass. By watching the glass Y at b 
it could be seen that no bubbles passed into the tube aaa, thus 
assuring a column of solid water. 
ATTACHMENT OF PIEZOMETERS. 
Probably no small portion of the discrepancy between results of 
tests on carrying capacity of various kinds of pipe is due to the method 
of measuring the pressure head. Apparatus of some form must be 
installed so that the pressure head will sustain a water or mercury 
column to an elevation corresponding to the true hydraulic gradient 
over the point of application of the pressure. In order to secure 
this result, all positive or negative influence of velocity head must 
