238 
Journal of Agricultural Research 
Vol. XXI, No. 4 
The difference as between progenies 2 and 1 was barely three times its 
prpbable error, while as between progenies 2 and 3 it was more than six 
times its probable error. 
Progenies were grown in 1919 from selfed seed of one plant in progeny 
No. 3 and of two plants in progeny No. 1 of 1918. In two of the 
progenies of 1919, flowers on a few plants were selfed and progenies from 
the resulting seed were grown in 1920. The percentages of 4-lock bolls 
on the plants of these progenies and on plants taken at random in a 
neighboring plot of bulk Pima, in both years, are given in Table VIII. 
Table VIII .—Percentages of 4-lock bolls in igig and ig20 in the progenies of plants 
selected for a high percentage of such bolls and in “bulk” plantings of Pima cotton 
Material. 
1919 
Material. 
, !920 
Num¬ 
ber of 
plants. 
Percentage of 
4-lock bolls. 
Num¬ 
ber of 
plants. 
Percentage of 
4-lock bolls. 
4-lock progenies: 
i -3 . 
. 
3“2 . 
Combined. . . . 
Bulk Pima. 
13 
20 
5 
38 
38 
8. 5±o. 65 
8. 7 ± -54 
to.g± .87 
9 - -37 
4 * 2± .27 
4-lock progenies: 
1-3-12 . 
3 - 2~4 . 
3 - 2 - 5 -.. 
Combined. 
Bulk Pima. 
45 
39 
42 
126 
50 
9. 2±o. 36 
22. 0± . 62 
23. 3± . 60 
17. 2± .30 
5 - 3 ± -29 
The descendants of plants which had been selected as showing a ten¬ 
dency to increased number of locks gave, in both years, significantly 
higher percentages of 4-lock bolls than plants taken at random in the 
general stock of the Pima variety. In 1919 the difference in favor of the 
4-lock progenies was 10 %, times and in 1920 it was 28 yi, times its prob¬ 
able error. In all three years (1918 to 1920) the descendants of selection 
No. 3 of 1917 gave a higher percentage of 4-lock bolls than the descendants 
of selection No. 1. These facts indicate clearly the heritable nature of 
this variation. 
There can be little doubt, therefore, that a strain having a materially 
higher percentage of 4-lock bolls could be developed by selection, although 
the practical importance of such increase in the Pima variety would 
seem to be small. Determination by Mr. W. G. Wells in 1919 of the 
mean weight of seed cotton per boll in 100 3-lock and 100 4-lock bolls 
showed that the latter averaged only 10 per cent heavier than the 
former. Similar determinations by Mr. W. F. Gilpin in 1920 on 50 bolls 
of each lock number showed a difference of only 7 per cent in favor of the 
4-lock bolls. Hence, even if it were possible to develop a strain of Pima 
cotton in which all the bolls would have 4 locks, the increase in yield of 
fiber and seed would not be likely to exceed 10 per cent 
