Peirce, L. C.
In 1956, parental, backcross, F1, and F2 populations of the cross Manalucie x Earlinorth were grown in a randomized block design. Progenies of F2 plants selected for superior yield, earliness, and fruit size were tested in 1957 with 20 F3 lines representing random F2 plants. In analyzing this material, two mettods were employed to estimate heritability for each character. The first, utilizing variance components of F2 and backcross populations, was based upon data of 1956 only. The second a regression analysis of F3 means on F2 performance reflected plant response in 1956 and 1957. Since a phenotype represents the response of genotype in a particular environment, a certain amount of bias might be expected if this environment were limited to one location in one year. Estimates of heritability based upon such data also would tend to show this bias. In contrast, values obtained under a wider sampling of environment generally reflect a reduction in interaction bias. The difference in heritability values shown in the following table for yield, earliness, and fruit size, respectively, therefore seems indicative of the relative effect of genotypic-environmental interaction in determining phenotype.
______________
* Journal Paper No. J-3533 of the Iowa Agricultural and Home Economics Experiment Station, Ames, Iowa. Project No. 1106.
______________________________________________
Heritability Heritability
(variance) (regression)
______________________________________________
Yield 39.4 % 27.l %
Earliness 38.7 % 49.3 %
Fruit size 75.5 % 68.1 %
______________________________________________
These results suggest that yield is especially sensitive to environment,
relative to earliness and fruit size. On this basis, F2 yield selections
might be of questionable value in a breeding program. The F3 results
revealed, however, that genetic improvement of this character exceeded that of
earliness in spite of lower heritability and greater interaction bias. Because
of this apparent contradiction, an attempt was made to determine the
enviromental sensitivity of genotypes in certain phenotypic classes. A study
of relative strain performance in successive years revealed strong year x
genotype interaction among plants of average yield, early ripening, and small
fruit size. Such interaction appeared to complicate selection of single,
early-ripening plants. In contrast, the reduced influence of interaction in
determining performance of superior yield-and fruit size genotypes undoubtedly
contributed toward improved selection efficiency for these characters.