THE FLOW OF WATER IN WOOD-STAVE PIPE, 21 
OPERATION OF GAUGES. 
From observations on the ground or study of the profile, the 
pressure in the pipe was known roughly. In assembling the gauge 
the proper number of iron-pipe units was installed in the long leg 
of the U so that the two ends of the mercury column balancing the 
pressure in the pipe would appear near the mid-point of the glass 
gauge tubes. 
The gauge, as assembled in figure 3, was set up beside the tapped 
pipe, the glass tubes being made truly vertical by means of a plumb 
bob. Shade was always provided for the gauges. 
With the cocks I and L open and G closed, mercury was filtered 
into the tube T. A paper-lined glass funnel was inserted at the top 
of T. The mercury filtering through a pin hole at the bottom of 
the paper funnel was thus cleansed at each experiment and the 
meniscus in each tube was made bright and clear. Mercury filtered 
into a tube of small diameter in this manner will fill up without air 
bubbles, but if it is poured into such a tube air bubbles will occupy 
long reaches of pipe and may not be found if they occur in the iron- 
pipe sections. When both legs of the gauge were filled to near the 
top of the lower glass, I was closed and G opened until the mercury 
column in M was pressed well down in the gauge, when G was again 
closed. Mercury was then added and G was opened from time to 
time until it might be allowed to remain open, the pressure holding 
the mercury column in sight on both glass gage legs. 
At this time there was probably a mixture of air and water above 
the mercury in M, but the air was driven out by alternately closing 
G and opening I for an instant. In using the gauges L was closed 
and G and I opened every few minutes, so that water and any accu- 
mulated air bubbles might be blown out of the pressure tube between 
the gauge and the wood pipe. 
Because of the use of unequal legs on the gauge there was danger 
of blowing mercury out of the gauge at I unless the cocks L and G 
were operated most carefully. To catch the mercury in the event 
of such an accident, the tube J was discharged into the bottle K 
which included a glass tube open to the air so that water was freely 
discharged but the mercury caught. 
Pulsations were nearly always present and as simultaneous readings 
of both low and high gauges were necessary in order to determine 
the length of the mercury column, readings were made in the follow- 
ing manner: 
Pulsation effect was reduced by partially closing either L or G 
until the mercury was barely " alive." This assured an average 
length of column. The cock was then completely closed, leaving a 
"dead" mercury column of the proper length. Both low and high 
gauges were carefully read to thousandths after which the cock was 
