24 BtJLLETll^ 1004, U. S. DEPARTMENT OF AGRICULTURE. 
These figures represent a high degree of correlation and point out 
that the quantity of water used by the crop certainly exerts a pre- 
ponderant influence on the yield. The fact that the quantity of 
water used correlates more closely with the total weight of crop 
produced than it does with the yield of either grain or straw is one 
that would be expected. It was not anticipated, however, that so 
close an agreement would exist between the two correlations of the 
yield of grain and water use and the yield of straw and water use as 
was shown by this study. It is, of course, necessary that there should 
be a yield of straw before a yield of grain can be produced, and it 
was thought probable that the total water use would correlate much 
more closely with the yield of straw than with the yield of grain or 
with the total weight. The fact that the three correlations agree 
very closely strikingly illustrates a statement previously made. This 
statement was that when affected by drought the wheat crop seems 
to spend its last energy in producing grain and that if there is any 
chance at all it will produce some yield of grain. This study indi- 
cates that a high yield of straw means a high yield of grain and that 
a low yield of straw means a low yield of grain. There have been 
a few cases when exceptionally favorable weather enabled wheat to 
fill so well that the yiem of grain was out of proportion to the yield 
of straw. These years are very infrequent, and as a whole the yields 
of grain and of straw are nearly proportional. 
It will be noted in Table 3 that in most years the precipitation 
has greatly exceeded the stored water used from the soil. In nearly 
every case a high precipitation means a high yield and a low precipi- 
tation a low yield. Very little relation is shown between the quan- 
tity of stored water used and the yield. There are two reasons why 
this is the case. In the first place the quantity of water stored in 
the soil is never enough to produce a good crop. Eainfall is neces- 
sary to supplement the soil moisture stored and a large quantity of 
stored moisture may be associated with a low rainfall, or vice versa. 
Another reason why the water used from the soil has little relation 
to the yield lies in the fact that the reduction of the soil water 
content is dependent upon the sufficiency or insufficiency of the rain- 
fall. In years of heavy precipitation there may be an actual increase 
in the quantity of stored water in the soil at harvest over the quantity 
present at seeding time, and in these years a negative correlation 
between the water used from the soil and the yield is shown. This 
condition has been observed, but is very infrequent in dry farming. 
That the stored water used from the soil actually plays an im- 
portant part in the yield of the crop was shown by correlating 
precipitation with yield. 
The correlation between the precipitation and the yield of grain 
was 0.74±0.03, that between the precipitation and the yield of 
straw was 0.60 ±0.04, and that between the precipitation and the 
total yield of gi^ain and straw was 0.73 ±0.03. These correlations 
are distinctly lower than those obtained in correlating the total 
water used and the yield. This makes it certain that though precipi- 
tation is the deciding factor in the yield of wheat the quantity of 
water stored in the soil in m^w cases is of great value in determining 
yields. 
That the quantity of stored water is of importance in determining 
yields was shown by another study. Comparison of two plats each 
