332 
Journal of Agricultural Research voi. xxvii. No. 6 
Table I.— Percentages of F x hybrids resulting from pollination of Egyptian (Pima) 
and of Upland (Lone Star and Acala) cotton flowers with approximately equal quanti¬ 
ties of Egyptian and Upland pollen 
Populations from double pollinations of: 
Number of — 
Percentage 
of hybrids. 
Plants. 
Hybrids. 
EXPERIMENT I 
i 
1 
! 
Pima with Pima+Lone Star and Lone Star+Pima... 
1,381 
442 
32. o±o. 85 
Lone Star with Lone Star+Pima and Pima-f-Lone Star 
752 
166 
22. I ± I. 02 
experiment 2 
♦ 
Pima with Pima-}-Acala and Acala-f Pima. 
968 
163 
16. 8±o. 81 
Acala with Acala+Pima and Pima-j-Acala. 
667 
220 
33. o±i. 23 
BOTH EXPERIMENTS AS ONE ARRAY ! 
1 
I 
Double pollinations on Egyptian.. 
i 
2,349 
605 
25. 8±o. 61 
Double pollinations on Upland. 
1,419 
386 
27. 2±o. 80 
Considering as one array all plants resulting from double-pollinations 
of the Egyptian (Pima) flowers and of the Upland flowers, respectively, 
(Table I, bottom section) it is evident that the percentages of hybrids 
were practically the same, the difference having amounted to only i .4 ± 1.0 
per cent. In both cases approximately three-quarters of the ovules were 
fertilized by pollen of the same type and the departures from the 50 per 
cent expected if there had been no selective fertilization are extremely 
significant. 
It is interesting that while in the first experiment the Pima flowers 
yielded a significantly higher percentage of hybrids than the Upland 
flowers (difference 9.9 ±1.33), in the second experiment it was the 
Upland flowers that yielded a significantly higher percentage of hybrids 
(difference 16.2 ± 1.47). The cause of this difference in the results of the 
two experiments may only be conjectured. Differences in the relative 
viability of the several pollens would seem to offer the most plausible 
explanation, but such data as were obtained show practically equal 
viability. 
Pollen from five flowers each of Pima and of Lone Star, each flower 
having been borne by a different plant, was tested as to its ejection when 
immersed in a 5 per cent solution of cane-sugar (7, p. 22-25). The 
estimated percentage of grains which ejected was 90 per cent for each of 
the Pima flowers and ranged from 80 to 90 per cent, with an average of 
87 per cent, for the five flowers of Lone Star. The tests were conducted 
during the period when the double pollinations of experiment 1 
were being made. Similar tests of pollen from four flowers of Acala 
cotton made at the outset of experiment 2, showed ejection of from 90 
to 95 per cent of the grains in the field of the microscope. 
SEQUENCE OF APPLICATION OF THE TWO POLLENS AS AFFECTING 
THE PERCENTAGE OF HYBRIDS 
Higher percentages of hybrids were obtained when the unlike pollen 
was applied first than when the like pollen was applied first. The data 
given in Table II show that this was the case for both types of cotton in 
