ADDITIVE AND NONADDITIVE GENETIC VARIANCES AND GENOTYPIC CORRELATIONS FOR YIELD AND OTHER TRAITS IN OATS
The oat cultivate 'Dorval', 'Kelsey', 'Stormont', 'Orbit', 'Goodfield', 'Tyler' and 'Egdolon' and two numbered lines were crossed in a 3 × 6 factorial design. F1's, F2's and bullied F2 progenies were grown in successive years at seeding rates of 2.7, 9.5 and at the commercial rare of 76 Kg/ha, respectively; the F1's in irrigated cages, the F2's and F3's in nonirrigated fields. Additive genetic variance was the most important component of the phenotypic variances among progenies. Major differences occurred between the F1's and F2's, but the F2's and F3's agreed closely. The percentages of additive variance (narrow sense heritability) for the joint F2, F3 analyses were: height (91); heading date (87); seed weight (74); yield per panicle (71); seeds per panicle (63); panicle number (58); stem diameter (55); plot yield (52). Important nonadditive variance was present in the F2, F3 data for plot yield (17%), stem diameter (15%) and seed weight (12%). All traits were positively correlated with each other in the F1's. Correlations were weaker in the F2 and F3 and reversed for panicle number. Seed weight showed the least correlation with other traits. These results are discussed from the practical viewpoint of combining strong straw with high grain yield. The usefulness of selecting for seed weight and panicle yield to improve plot yield is underlined.