USE OF WATER BY SPRIXG WHEAT OX GREAT PLAINS. \) 
The correlation between the total quantity of water used per day 
and that portion of this quantity that is obtained by reduction of the 
water content of the soil (designated daily use from soil) is 0.48 ±: 
0.07 for plat A and 0.5T±0.06 for plat C or D. The correlation be- 
tween the total quantity used per day and that portion of this quan- 
tity that is supplied bv precipitation during the period under study 
is 0.30±0.08 for plat A and 0.29±0.08 for plat C or D. Both these 
are positive correlations, but the correlation of the total quantity 
with the precipitation is of a low order. The higher correlation of 
the total with the portion obtained from reduction of the soil water 
content indicates that under the conditions specified for this study 
the stored water is a more dependable source of supply than the 
precipitation. 
The correlation between the two parts that make up the total daily 
rate of use (the rate of use from the soil and the precipitation) is a 
negative one of — 0.73±0.04 for plat A and — 0.67d=0.08 for plat C 
or D. This is a rather high correlation and expresses the high degree 
to which the two parts that make up the whole are complementary, 
as previously explained. 
Table 1 shows the rates of the use of water per day varying in indi- 
vidual years at individual stations from 0.10 to 0.33 inch per day. 
That these differences are largely real differences and not experi- 
mental error is indicated by the agreement between the two plats. 
The rate of use may not be the same on the two plats in any year, 
because they may be supporting different amounts of vegetation, but 
changes from year to year are in nearly every case in the same direc- 
tion on the two plats. The rate of water use for any year is undoubt- 
edly not correct to within 0.01 inch and in some cases the error may 
be several hundredths of an inch, but the use of as lon^ periods as 
possible for each year and the inclusion of as many years as possible 
in the data should make the average results for a station reliable. 
The 53 station years studied show an average rate of use of 0.17 
inch per day on plat A with an average yield of 17 bushels per acre, 
and an average rate of 0.19 inch per day on plat C or D with an aver- 
age yield of 22.4 bushels per acre. In some cases the rainfall during 
the period under study has been greater than the quantity of water 
used, meaning that the quantity of water in the soil increased, and 
in other cases the rainfall has contributed as little as 0.01 inch per 
day to the total. The average rate of rainfall was 0.09 inch per day. 
Some points may be shown more clearly by assembling the average 
rates of use per day for each station in a separate table. This is 
done in Table 2, which gives the average rates of use per day for 
both plats A and C or D at each station. In this table, as in Table 1, 
the stations are arranged as nearly as may be with respect to their 
geographical location, the more northern ones at the top of the table 
and the more southern ones at the bottom. This table shows im- 
mediately for plat A the dependence of the rate upon the geo- 
graphical location. The northern stations show in general lower 
rates of use than the southern. Exceptions are to be noted in the 
case of the record of a single year at Ardmore, which shows a higher 
rate than its location would indicate, and a single year at Garden 
City, which has a low rate. This low rate was made by a very light 
crop and should not therefore be considered as an exception. The 
21511*'— 23 2 
