Feb. 9,1924 
Selective Fertilization in Cotton 
335 
developed from the double-pollinated Pima flowers in the experiments 
here described was 16.8 ± o. 14. Therefore, an average of 4.7 ovules had 
failed to develop into seeds. The proportion of hybrids in the adult popu¬ 
lation from seeds borne by Pima plants, taking the populations of the 
two experiments as one array, was 25.8 per cent (Table I). It is com¬ 
puted therefore that the mean number of heterozygous seeds per boll was 
4*33 ( 2 5-8 per cent of 16.8). Assuming that all of the ovules which failed 
to develop represent heterozygotes or uncompleted heterozygous unions 
and adding the mean number of undeveloped ovules to the mean number 
of heterozygous seeds, a total of 9.03 is obtained as representing the mean 
number of possible heterozygotes per boll. Even this number is only 42 
per cent of the mean number of ovules. A similar computation in regard 
to the Upland varieties in these experiments is impracticable because 
satisfactory data as to their mean numbers of ovules are not available. 
There is, however, no reason to assume that all or most of the unde¬ 
veloped ovules represent unsuccessful unions or attempts at union with 
unlike male gametes. The mean number of seeds per boll from the double- 
pollinated Pima flowers of these experiments (16.8) exceeds the average 
(16.5), for 10 lots of bolls from naturally pollinated flowers of the same 
variety, as given in another publication (7, p. 51, Table 30). The highest 
mean number for any of these lots was 18.6 or 2.9 fewer than the mean 
number of ovules. It is evident that under the most favorable condi¬ 
tions a number of the ovules fail to develop, probably because they are 
defective or because they are not reached by pollen tubes. The conclu¬ 
sions seem warranted, therefore, that the percentage of undeveloped 
ovules in the present experiments was not abnormally high and that there 
had been no selective survival at fertilization or in the early stages of 
development of the zygote. 
RATE OF GROWTH OF THE POLLEN TUBES IN RELATION TO SELECTIVE 
FERTILIZATION 
In seeking an explanation of selective fertilization, the possibility of a 
difference in the rapidity of germination and of pollen-tube development 
of the like and unlike pollens is the first point to be considered. An 
experiment was described in another paper (7, p. 42) in which it was 
sought, by excising the stigmas and style at successive intervals of time 
after pollination, and by comparing the degrees of fertilization thus at¬ 
tained, to determine the relative rates of growth of Pima and of Upland 
pollen deposited separately on the stigmas of different Pima flowers. 
The growth rates of the like and unlike pollens, under these conditions, 
were found to be very similar. The experiment was repeated in 1923, a 
number of Pima flower buds having been emasculated the evening before 
anthesis and pollinated at 8 a. m. the next day, some of the flowers with 
Pima pollen and others with Upland pollen (Acala variety). The styles 
of equal numbers of flowers of each pollination were then severed at the 
summit of the ovary at intervals of 8, 10, and 12 hours after the pollen 
was deposited.® A record was kept of the number of bolls which devel¬ 
oped from each lot of flowers and of the number of seeds in each boll. 
The results are summarized in Table IV. 
8 This method was used by Heribert-Nilsson (2) in an investigation of the rate of pollen-tube growth in 
Oenothera. His results showed a more rapid growth of the like pollen. 
