Feb. 9, 1924 
Selective Fertilization in Cotton 
33i 
flowers were therefore mixed thoroughly and a representative sample of 
each lot was planted. The quantities of seed from the several pollina¬ 
tions of the Upland flowers were so small that all were planted. 
The seeds representing each sequence of pollination (like -f unlike and 
unlike -f like) on each type of cotton in each experiment were planted as 
separate populations and these populations were further subdivided, 
seeds from the upper and the lower halves of the bolls having been 
planted separately. This gave a total of eight ultimate populations 
from each of the two experiments, as follows: 
Pima X (Pima -f Upland) 
Do. 
PimaX(Upland-}-Pima). .. 
Uplanc^X (Upland-}- Pima). 
Do. 
Upland X (Pima+Upland). 
Do. 
Upper halves of bolls. 
Lower halves of bolls. 
Upper halves of bolls. 
Lower halves of bolls. 
Upper halves of bolls. 
Lower halves of bolls. 
Upper halves of bolls. 
Lower halves of bolls. 
These ultimate populations were variously combined in considering 
the data from different points of view. The plants grown from upper 
and lower seeds of the bolls from each pollination are treated as separate 
populations only under the heading “Rate of growth of the pollen tubes 
in relation to selective fertilization. ” 
Four seeds were planted to the hill and no thinning was done. A 
total of 2,349 plants was obtained from the double-pollinated flowers 
of Pima and a total of 1,419 plants from the double-pollinated flowers 
of the Upland varieties. The difference in number of plants was due 
to the quantities of seed of the two types available for planting and not 
to better survival in the germinating and seedling stages of the plants 
from seeds produced by Pima flowers. For both types of cotton the 
number of plants which survived averaged 2.6 per hill. 
On July 19 and 20, when the plants had developed sufficiently to make 
recognition of the hybrids unquestionable, 4 counts were made of the 
total number of plants and of Fj hybrids in each population and from 
these data the percentages of hybrids were computed. 
COMPARISON OF EGYPTIAN AND UPLAND COTTONS AS TO SELEC¬ 
TIVE FERTILIZATION 
The data from the two experiments, showing the evidence for selective 
fertilization in both types of cotton, are presented in Table I. In each 
case the results from pollination first with like and then with unlike 
pollen, and the converse, are combined. It will be shown that the 
pollen applied first fertilized a greater number of ovules than the pollen 
applied afterward to the same stigmas, but it seems reasonable to assume 
that the percentage of hybrids obtained by combining the populations 
from like-funlike and unlike + like pollinations as one array corresponds 
with what would have been obtained had it been practicable to mix 
the two pollens in equal quantity before application to the stigmas. 5 
4 The characters distinguishing Fi hybrids of Pima and Upland cotton from the parental types were 
described in another paper (8, p. 6, 7). 
5 Nevertheless, a small error is involved owing to the fact that populations of unequal size were obtained 
from the pollinations like+unlike and unlike-Hike. If, however, the percentages of hybrids for each 
sequence of pollination (as given in Table II) are averaged for each variety in each experiment, the aver¬ 
ages depart but slightly from the percentages based upon the two sequences taken as one array, as given 
in Table I, and the significance of the results is not altered. 
