U. S. DEPARTMENT OF AGRICULTURE. 
number of basal branches, and the selection work has always aimed to 
reduce this branching habit. The manner of branching can best 
be studied in open drill rows where the plants, thinned down to one 
to the inch, are not crowded and can develop normally. In open drill 
rows the check selection never develops more than two basal branches, 
and with more than half of the plants it has none at all. Any selec- 
tion having more than two basal branches was discarded. 
Seed vitality received more attention as the work progressed. In 
addition to making germination test's and taking notes on the stand 
secured in the field, a study was made of the number of seeds per boll 
of all promising selections. The number of seeds per boll when 
complete fertilization occurs is 10. This number is rarely approached 
under field conditions. The data of 1919 and 1920 confirm the earlier 
work. The average number of seeds per boll is an inherited character 
directly associated with the vitality of the seed. For example, Selec- 
tion 1923, with the best vitality of seed, has a high count of seeds per 
boll, and Selection 1812, with poor vitality of seed, has a low count of 
seeds per boll. Hence, all selections having a decidedly low count 
of seeds per boll were eliminated unless the count for the check selec- 
tion, grown in the same part of the field, also ran very low. 
In the years 1914 and 1915 marked differences in resistance to wilt 
were noted in fiber-flax selections when a natural, survival of the more 
resistant selections took place. The check, Selection'No. 5, was at this 
time superior to most of the others in resistance to wilt. In 1919 the 
selection plat at Croswell, Mich., was planted on soil where flax stems 
had been previously spread. All selections from the vYhite-Blossom 
Dutch variety were completely destroyed by flax wilt, and not 
more than one -fourth of the plants among the selections from the 
Blue-Blossom Dutch variety, the other commercial fiber flax, sur- 
vived. The check selection, planted in every other row. was more 
than 85 per cent resistant. Corroboration of the results was secured 
in 1920 through the cooperation of the Office of Cereal Investigations 
of the Bureau of Plant Industry. Plats were sown with seeds fur- 
nished from the selection plats of fiber flax by Mr. J. C. Brinsmade, 
jr., at Mandan and by Mr. W. E. Brentzel at Fargo. X. Dak. In 
Mandan the check, Selection Xo. 5, was over 80 per cent resistant to 
wilt, while the Blue-Blossom Dutch variety was 21 per cent and the 
\Yhite-Blossom Dutch only 1 per cent resistant. At Fargo, Selec- 
tion Xo. 5 compared favorably with the most resistant seed flaxes in 
both resistance to wilt and seed yield. At both places the order of 
resistance of the different fiber -flax selections was approximately the 
same as that obtained in the selection plat in Michigan ; those selec- 
tions which were more resistant in Michigan were also the ones most 
resistant in North Dakota. This work makes it possible to eliminate 
