330 
Journal of Agricultural Research voi. xxvii, No. 6 
Selective fertilization, in favor of like as compared with unlike, although 
compatible pollen when both are present on the stigmas, has been shown 
by Jones to occur in maize and tomato (j, 4, 5, 6). Experiments of 
Heribert-Nilsson indicated that it takes place also in Oenothera ( 2 ). The 
fact that members of such distantly related families as Poaceae, Mal¬ 
vaceae, Onagraceae, and Solanaceae manifest selective fertilization in 
favor of the like pollen, points to the conclusion that the phenomenon is 
by no means uncommon among flowering plants. Its probable impor¬ 
tance as a factor contributing to the differentiation of plant forms has 
been discussed by Jones (5). 
The evidence for selective fertilization brought forward by Jones, 
supplemented by data given in this paper, applies only to cases when 
“germ cells from two individuals of different type are presented at the 
same time in excess so that not all can fulfill their function’* (4, p. 253), 
and has no bearing upon the question whether there is selective fertili¬ 
zation among the several gametes produced by the same individual. 
METHOD OF INVESTIGATION 
The experiments were conducted at the Cooperative Testing Station 
at Sacaton, Arizona. The pollinations were made in 1922. The flowers 
to be pollinated were emasculated in the evening preceding anthesis 
and were enclosed in paper bags to prevent accidental cross-pollination. 
The flowers which were to supply the pollen were bagged at the same time. 
The pollinations were made at 1 p. m. of the following day, the procedure 
having been to detach the flower supplying the pollen and brush the 
stigmas of the emasculated flower lightly with the clustered anthers 
of the staminate flower (7, PL /, IV, V). 
Because of the mechanical difficulty of mixing the two kinds of pollen 
in equal proportion, recourse was had to the method of double pollina¬ 
tion used in an earlier experiment (7, p, 46). This consisted in pollinating 
one-half of the flowers first with the like pollen and then with the unlike 
pollen and reversing the order of application on an equal number of 
flowers. The two pollens were applied in as nearly as possible equal 
quantity and the interval between the two pollinations was negligible. 
Two experiments were performed, the pollinations of the first experi¬ 
ment having been made during the period July 21 to 30 and those of the 
second experiment August 10 to 13. The Egyptian type of cotton was 
represented in both experiments by the Pima variety. The Upland 
type was represented in the first experiment by the Lone Star variety 
and in the second experiment by the Acala variety. The stocks of each 
variety used in the experiments were not inbred families but had been 
grown from seed produced under conditions of isolation and were judged 
to be varietally pure. Well grown plants were used in both experiments, 
those in experiment 1 having been located in ad jacent plats, and those in 
experiment 2 in the same plat. Two hundred flowers of each variety 
were pollinated in each experiment. 
The seeds produced by the double-pollinated flowers were planted in 
1923. As many more bolls matured from the double-pollinated flowers 
on the Pima than on the Upland plants, the former yielded much greater 
quantities of seed than the latter. 3 The seeds from each lot of Pima 
* Abscission of the young bolls is always more pronounced in Upland than in Pima cotton as grown 
Sacaton. 
