20 CIRCULAR NO. 115, BUREAU OF PLANT INDUSTRY. 
Until the habits of branching are taken into account it seems 
impossible to explain the widely different results that are often 
secured when the same experiments are repeated in different places 
or in the same place in different seasons. From the present point 
of view it is easy to understand that merely statistical experiments 
made without recognizing the effects of different methods of thin- 
ning upon the formation of branches would be likely to reach 
only ambiguous results. The development of the branches, though 
very easily influenced in the early stages of growth, completely alters 
the subsequent behavior of the plants. The effect seems out of all 
proportion to the exciting cause, hke touching off a charge of powder 
or pulling the trigger of a gun. 
Wider spacing appears as the only alternative as long as the young 
plants are led to put forth a full equipment of vegetative limbs by 
too much exposure in the early stages of growth. That the develop- 
ment of these limbs may be avoided by a later and more gradual 
thinning of the young plants must be recognized before it is possible 
to understand the advantages of the new system. When good crops 
are produced on rows that are not thinned at all, it is because the 
plants remain so close together that no vegetative limbs are devel- 
oped. The new system provides for a more regular and effective ap- 
plication of the same principle of suppression of vegetable branches. 
THINNING EXPERIMENTS WITH DURANGO COTTON. 
The behavior of the Durango cotton at Norfolk, Va., in the season 
of 1912 affords an excellent illustration of the application of the 
principle of controlling the formation of branches as a means of 
securing earlier and larger crops of cotton under short-season condi- 
tions. Planted in a row test with other varieties and thinned in the 
usual manner to the ordinary distances, the Durango cotton, which 
is unusually productive for a long-staple variety, yielded at the rate 
of about 1,175 pounds of seed cotton per acre, though considerably 
exceeded by the Trice, an extra-early short-staple variety, which 
produced at the rate of 1,756 pounds per acre. In a field planting 
on lighter and less fertile soil the rows of the Durango cotton that 
were thinned in the usual manner to ordinary distances yielded at 
an average rate of 909 pounds of seed cotton per acre, while alter- 
nate rows that were thinned late and left with the piants closer 
together yielded at a rate of 1,391 pounds, or about 53 per cent 
higher than the others. Table I shows some of the figures obtained 
from the Norfolk experiment by Mr. G. S. Meloy. 
[Cir. 115] 
