CEREAL EXPERIMENTS AT DICKINSON, N. DAK. 39 
VARIETAL TESTS. 
A much larger number of varieties is tested each year in the 
nursery than in the field plats. The new varieties which are con- 
tinually being introduced and originated are here given their first 
trial. The land used is alternately cropped and fallowed. The seed 
is sown by hand in a furrow about 3 inches in depth, which has 
been opened with a wheel hoe. It is then covered by a special at- 
tachment upon the same tool. The rows are 17 feet long and 12 
inches apart. Every third row is planted to the same variety, as a 
check. Careful notes are taken throughout the season on the com- 
parative behavior of the different varieties. At harvest time each 
row is cut (a part.of the row is pulled) separately, labeled, and the 
Fie. 7.—General view of the cereal nursery at the Dickinson substation in 1918. 
heads covered with paper sacks to prevent mixture or damage by 
birds. The total weight of each row and the total weight of grain 
per row are obtained at thrashing time. In addition, the following 
notes are taken on 25 plants from each lot: Height of plant, length 
of head, number of culms per plant, number of heads per plant, total 
weight of plants, and weight of grain. These notes for a series of 
years give a rather accurate analysis of each variety. The correla- 
tion of the different characters, with yield, is the primary object of 
this analysis. : 
The weight of grain per row gives the relative yielding power of 
the different varieties. Computed yields are determined from the 
use of the check rows. New and untried varieties in this way are 
compared with standard and better known varieties. Any variety 
producing a relatively high yield for several years is increased and 
