49 2 
Journal of Agricultural Research 
Vol. XXI, No. 7 
Inspection of the yields from tlie various rates of seeding at Morris, 
with no border rows removed, show that the 8o-pound seeding gave the 
highest yield and the others ranked in the following order: 96 pounds, 
112 pounds, 64 pounds, and 48 pounds. This rank is not changed by 
the removal of one or two border rows. The least significant difference 
between the yields of any two rates of seeding was found for this test by 
using the formula 
standard deviation X 0.6745 
Jn 
where n denotes the number of plots for each rate ( 22 ). The result was 
multiplied by 3.2, which makes the odds 30 to 1 that the result is not 
due to chance ( 16 ). For no border rows removed, this figure is 4.93. 
Where one and two border rows were removed, the figures are 4.96 and 
6.27, respectively. If these are used in a broad way in the interpre¬ 
tation of the results, it appears that where 110 border rows were removed 
the 8o-pound rate of seeding gave better results than the 64-pound 
rate, which in turn gave better results than the 48-pound rate. The 
interpretation of the results where one and two border rows were 
removed would not be much different. 
At University Farm the rank with one and two border rows removed 
tends to be somewhat different than with no borders removed. Employ¬ 
ing in a broad way 4.51, 4.06, and 4.38 bushels, respectively, as the 
least significant differences where none, one, and two border rows were 
removed, an interpretation of the results may be made. Where no 
border rows were removed, the 8o-pound rate appears to have yielded 
the highest, with the 96-pound and 112-pound rates equal and in turn 
higher than the two lower rates. For the tests with one and two 
border rows removed, the 96-pound rate seems to have maintained a 
lead over the 80- and 112-pound rate and to have yielded significantly 
higher than the 64- or 48-pound rate of seeding. 
The results for the variety tests of oats, wheat, and barley in 1918 
with none, one, two, and three border rows removed are given in Table 
VII. 
For the oat and barley varieties there are some changes in rank 
brought about by the removal of border rows. It is of interest to note 
that the removal of borders makes no change in the rank of the last 
four varieties, which were also the lowest in rank in the 1917 tests. In 
the barley variety test the rank of the Manchuria cross and the Chevalier 
are not changed. The Chevalier ranked the highest in 1917. 
With the same formula as was used in the rate of seeding tests to 
secure figures representing the least significant differences, the follow¬ 
ing were secured for the oat tests: No borders removed, 5.26 bushels; 
one removed, 4.54 bushels; two removed, 4.46 bushels; and three re¬ 
moved, 6.90 bushels. These amounts have been subtracted from the 
