30 BULLETIN 1004, U. S. DEPAKTMENT OF AGRlCULTXJEE. 
It has not been possible to compute yields by the linear- regression 
formula as closely at North Platte as at Edgeley. This may be 
because there is a greater experimental error at North Platte than 
at Edgeley, for it is necessary to include the soil to a depth of 6 
feet in the samplings. It is probable, however, that the real reason 
lies in the fact that North Platte lies in a section not so well adapted 
to spring wheat as Edgeley. Extreme conditions of temperature 
that affect wheat adversely are much more common. The lack of 
the same correlation at North Platte as at Edgeley may and probably 
does indicate the greater proportional effect that conditions other 
than the quantity of water used have upon the yield. 
GENERAL DISCUSSION OF THE RESULTS. 
Physiological studies of the use of soil water by crop plants have 
determined the use of water by the plants alone. Such studies have 
been of utmost importance in their contribution to knowledge of 
the relation between the plant and its environment and between 
different groups of plants. As such studies are made under con- 
ditions that permit the escape of no water except through the plant 
they measure the maximum possibilities of growth or production 
from a given quantity of water. 
The present study attempts to measure the rate of use and the 
quantities of water used under the complex of conditions that environ 
the crop in the field. The data available for study are of such a char- 
acter that only approximations for given conditions and averages can 
be arrived at. Indeed, the problem itself is not of a nature to permit 
exact determinations. If the methods were so refined and the work 
so replicated as to approach exactness in the results for any given 
condition, the complex of this condition might never be duplicated. 
To furnish usable information marking a distinct advance in the 
knowledge of the subject it is not necessary that determinations of the 
rate of use or the^ quantity of water used be exact. It is not of pri- 
mary importance to know whether a crop uses water at the rate of 0.15 
or 0.16 inch per day, but it is important to know whether the rate 
be 0.15, 0.30, 0.60 inch, or even more per day. The wider variations 
and greater ranges are of primary agricultural importance, and 
within such ranges the approximations of rates and quantities arrived 
at meet the requirements made of them. 
Many writers draw erroneous conclusions of the possibilities of a 
given rainfall by not recognizing the fact that yield is not directly 
proportional to the quantity of water used. A common presentation 
is that if a given number of inches of water produces a certain size of 
crop then double the quantity of water will produce twice that much. 
The data here presented show that such calculations may be very mis- 
leading. An analogy may be drawn from the animal world to illus- 
trate the source of error in such calculations. In feeding live stock 
it is well recognized that a certain quantity of feed is consumed as a 
living ration. This quantity alone results in zero gTowth, and some 
quantity in addition to this must be supplied if there is to be an incre- 
ment of growth. The results of the present study show that in the 
field the consumption of a certain quantity of water is necessary be- 
fore any yield of grain is produced. Or, stated in another way, the 
consumption of a certain quantity of water only results in a zero yield. 
This quantity may vary in the Great Plains from perhaps 4 to 10 
