CEREAL INVESTIGATIONS ON THE BELLE FOURCHE FARM. 11] 
acre in area. The alleys between these plats have been 19.2 inches 
in width. By the use of plats and alleys of these dimensions it was 
possible to sow five plats within the area formerly occupied by one 
tenth-acre plat. As the plants draw considerable moisture and 
plant food from the alleys, it has been thought fair to consider these 
one-fifty-fifth-acre plats as one-fiftieth-acre plats in computing acre 
yields. 
REPLICATION OF PLATS. 
In 1908, 1909, and 1910, when the tests were conducted on tenth- 
acre plats, there was only a single plat of each variety. Check plats 
of standard varieties of each cereal were sown at regular intervals 
m 1909 and 1910, and in most of the tests in 1908. As this method 
did not appear to be entirely satisfactory, a change was made in 
1911 mm some of the tests and in all those conducted in 1912 and 
1913. The size of the plats was reduced, as stated in the preceding 
paragraph, and the tests were replicated. In the varietal tests, five 
plats of each variety were grown. In rate-of-seeding and date-of- 
seeding experiments it has been considered sufficient to grow three 
plats of each rate or date, as there is a correlation between the differ- 
ent parts of the experiment which is not found in the varietal tests. 
PREPARATION OF THE LAND. 
In preparing the land for experimental work the aim has been to 
keep within practical farm limits as far as possible. The plowing has 
been done at a moderate depth, 6 to 8 inches, and subsequent treat- 
ment has been in accord with the best farm practice. The necessity 
for keeping the land uniform has sometimes required hand work, 
such as the removal of weeds, and probably more cultivation at 
times than would be done by a practical farmer. The work also has 
been more timely than is usually the case on large farms. 
Most of the experimental work has been on land fallowed during 
the preceding year. Fallowing also has been necessary to insure 
uniformity in soil conditions. When the experimental work was 
begun, this appeared to be the most practicable method of producing 
-erops under dry-land conditions. 
The usual practice has been to plow in the fall after the crop was 
removed, or the following spring if conditions were not favorable 
for fall plowing. Ground plowed in the fall has been left rough or 
has been worked dotvn with the disk and spike-tooth harrows, ac- 
cording to the moisture condition. If it contained considerable 
moisture it was worked; otherwise it was not. All fallow has been 
worked when necessary to remove weeds or to prevent evaporation. 
The spring-tooth harrow has proved to be the most satisfactory tool 
_ for working fallow land. Other implements, such as the smoothing 
