22 BULLETIN 376, U. S. DEPARTMENT OF AGRICULTURE. 
opened. This process was repeated every 10 minutes. All other 
readings were taken by alternately reading high and low gauges 
with the mercury just alive, the corresponding reading for the other 
gauge being computed from the dead readings as described above. 
Since the only change in the total length of mercury thread was 
due to temperature changes, and since the gauges, which were made 
of the highest grade of manometer tubing, were practically uniform 
in diameter, no error was introduced by reading but one leg at a 
time, alternately. (See p. 93.) 
DETERMINATION OF LOST HEAD. 
The exact amount of h f (fig. 1) must be determined. Where a 
water column is used, say at gauge No. 2, the elevation E 2 is the 
gauge reading added to the elevation of the gauge zero above an 
assumed datum, with proper corrections (see p. 23). Where a 
mercury manometer of the U-tube pattern is used, the reasoning is 
as follows: It is desired to know the elevation E x (fig. 1) for a water 
column which is the equivalent of a mercury column in a U-tube 
placed as for gauge No. 1. Referring to figure 2, the mercury in 
the two legs of the U-tube below c-d will be seen to balance. There- 
fore the pressure of the water at c is just balanced by the column of 
mercury X. But the pressure at c equals that at d. If the mercury 
X were replaced with water it would reach an elevation sX above 
d, where s is the specific gravity of the particular mercury in the 
gauge, compared with the particular water in the pipe. But the 
elevation to which this water column would reach is the desired 
elevation, E x . Therefore the elevation E x =sX + y above the as- 
sumed datum. As applied to these experiments, referring to figures 
1 and 2, the difference in elevation between the readings of the low 
gauge and the high gauge multiplied by the specific gravity of the 
mercury and added to the elevation of the low-gauge reading gave 
the elevation of the equivalent water column when the proper cor- 
rections had been applied. 
-*- • CORRECTIONS. 
Although quite numerous, the principles involved in all of the 
necessary corrections have been the subjects of such thorough inves- 
tigation that appreciable errors are not liable to result from their use. 
Temperature. — Corrections are necessary for the temperature 
changes in both air and water. A temperature of 15° C. was adopted 
as standard and the specific gravity of the mercury used in the tests 
was referred to that temperature, being compared to distilled water 
at the same temperature. 
The mercury column balances the pressure of the water in the 
pipe, but this water may be either heavier or lighter than distilled 
