363] B. E. Invingston 165 
spiration but to measure the atmometric index of the air in 
terms of its effect upon a standard instrument whose internal 
conditions do not change. The internal conditions of each 
plant or group of plants must be studied in relation to the un- 
changing ones of the instrument. A given temperature change 
does not affect all objects or processes alike, yet we do not con- 
struct a new thermometer scale for each object or process with 
which we deal. It may be well to mention in this connection 
that atmometry should furnish climatological data applicable 
to many fields of endeavor; the animal ecologist requires these 
data as much as does the plant ecologist, and irrigation en- 
gineers and students of atmospheric hygiene and ventilation 
all have use for atmometric measurements. 
In choosing the instrument to be used the first condition 
to be met is that its internal conditions or characteristics 
should not alter; they should be uninfluenced by changes 
in the surroundings, for it is changes in the latter that we wish 
to measure. This requirement immediately excludes all forms 
of free water surfaces, since they alter with wind, etc. Never- 
theless, since an open pan of water is the from of atmometer 
employed by the U. 8S. Weather Bureau, since this is the sim- 
plest form of instrument that is useful in any way, and since 
data obtained with this pan will surely prove of much greater 
value than no atmometric data at all, the pan of water must be 
accepted as the crudest and most imperfect form of atmome- 
ter. It should be added that if pans of water are used they 
should generally be of the same form, size, etc., as the stand- 
ard recently adopted by the U. 8S. Weather Bureau. If this be 
adhered to, all pan measurements will be comparable among 
themselves and with the Weather Bureau data, as far as this is 
possible with that general class of instruments. 
The second requirement for an evaporating surface is that 
it should be as sensitive to all the effective conditions of the 
surroundings as is possible, without any alteration in its in- 
ternal characteristics. It should therefore be a surface that 
is freely exposed to wind action. A nearly ideal surface would 
