RESEARCH METHODS IX STUDY OF FOREST ENVIRONMENT. 159 
higher than in A', the difference in the height of the two columns depending 
upon the height of the cup above the water level in the reservoir. The mer- 
cury in B B' is drawn into the bulb C, where water rising from the reservoir 
is freely allowed to pass, supplying in the usual way the water lost from the 
surface of the cup. The mercury columns in the tube A A' and B B' remain 
in equilibrium in the position indicated 'n the figure, so long as the water 
loss from the surface of the cup by evaporation equals or exceeds the absorp- 
tion from without by any part of its surface. In t ? mes of rain, when the water 
loss from the surface of the cup by evaporation is less than the absorption from 
without, the automatic mercury valves become reversed. The mercury col- 
umn falls in A and rises in A' at the same time that the mercury in bulb C 
drops into the tube below and rises in the tube B' (the height to which the 
mercury rises in this tube depending upon the height of the cup above the- 
water level in the reservoir), thus effectually preventing water from entering 
the reservoir from this direction. The readings obtained give the actual evapo- 
ration minus the error introduced by the volume change required for the op- 
eration of the mercury valves, the value of the error thus introduced depending 
upon the number of complete reversals of the valves. 
Standardization. 
The Livingston porous-cup atmometers, whether to be used with 
or without the Shive nonabsorbing apparatus, should be obtained in 
the standardized form, which insures comparability of the results 
obtained with different instruments and by different investigators. 
This standardization of the instruments is one of the strongest fea- 
tures, making it possible, for the first time, for different investi- 
gators to speak of evaporation in common terms. Since the stand- 
ardizing can not be readily accomplished in the field, it is suffi- 
cient to state that it is very carefully done at the Johns Hopkins 
laboratory, where each new instrument is compared with one or 
more tried instruments. Assuming the rate of evaporation for a cer- 
tain time to be 100 in a standard instrument, the coefficient of an in- 
strument which in the same period evaporated 150 units, would be 
0.67. In short, the coefficient given, for any instrument which has 
been " standardized," is the amount by which all records of that 
instrument should be multiplied, to reduce them to a standard basis. 
Computation of field results. 
In computing the field results to a standard basis, all cups used for 
a few days or weeks are restandardized. Provided the new evapor- 
ating coefficient is found to differ from the original, the probable 
average coefficient during regular cycles or periods is calculated. 
It has been shown experimentally that the change in the evaporation 
coefficient takes place gradually and uniformly during the period of 
operation (159). Thus, assume that a given atmometer cup having 
an initial coefficient of 0.62 was exposed in a habitat for a period of 
four weeks during which it was read three times; namely, at the 
