C orn I nves tig a tions 
55 
A' COMPARISON OF FIRST, SECOND, AND THIRD GENERATION PURE LINE 
HYBRIDS 
In 1915 several F 1 plants in each of the eight combinations 
shown in Table 16 were fertilized with composite pollen of 15 
sister plants, in order to produce seed for second generation 
(F 2 ) hybrid plants to be tested in 1916. In 1916, F 2 hybrid 
seed was again produced in the same manner, and third genera- 
tion (F 3 ) hybrid seed was similarly produced from F 2 plants. 
First generation hybrid seed was produced anew each year, for 
comparison. The comparative 1916 results for the individual 
first and second generation hybrids are shown in Table 16, in 
contrast with the average results for the pure lines and the 
original corn. The original corn yielded 37.5 bushels per acre; 
the seven pure lines averaged 9.8 bushels ; the eight first genera- 
tion hybrids averaged 51.7 bushels; and the second generation 
hybrids 25.1 bushels per acre. Thus the F t crop yielded 138 
per cent as much, the F 2 crop 67 per cent as much, and the pure 
lines 26 per cent as much as the original variety from which 
they were produced. 
The increased degree of homozygosity due to Mendelian 
segregation and recombination of parental factors would ac- 
count for the reduction in yield of the second generation plants, 
which is nearly in accordance with expectancy in a hybrid popu- 
lation in which the factors represented were all derived from 
two homozygous individuals. 
In 1917, F 15 F 2 , and F q pure line hybrids were available 
for comparison. (Table 17.) The original seed yielded 46 
bushels, the F n averaged 51.5 bushels, the F 2 29.4 bushels, and 
the F ; , 25.6 bushels per acre. This extreme reduction in yield 
of the F 2 and F 3 crop below the yields of the F, and the orig- 
inal variety is evidence to show that, in any application of the 
principles of corn improvement thru crossing two pure lines, it 
is essential to avoid selecting seed from the hybrid progeny. An 
experiment is under way in which seed of a large number of 
elemental strains are mixed to be grown thereafter as’ a variety 
without further controlled inbreeding and hybridization. By 
mixing a large number of strains, the chance of union of like 
gametes is reduced and when enough strains are used will be- 
come small, as in ordinary commercial varieties. At the same 
time, advantage would be taken of the process of weeding out 
inferior strains thru inbreeding. 
The two years’ results with F, and F 2 crop are summarized 
in Table 18 and show respective yields of 52.2 and 27.8 bushels 
per acre in comparison with 41.7 bushels for the original corn. 
