ESTIMATION OF ADDITIVE AND EPISTATIC GENETIC VARIANCES FOR AGRONOMIC TRAITS IN A POPULATION OF DOUBLED-HAPLOID LINES OF WHEAT

In order to determine a selection strategy for a population of winter wheal subjected to recurrent selection, we assessed the relative exten t of both additive and epistatic effects for agronomic traits involved in yield performance. The partitioning (between and within mother pla nt) of the genetic variance in doubled-haploid lines derived from the intercrossed population provided estimates of additive and epistatic a dditive x additive variances at the pure line level. Two similar exper iments with 56 lines in 1992 and with 73 in 1993 were conducted at Gif sur Yvette in France. Results showed that 1993 was less favourable fo r yield performance than 1992. Even when genotype-by-year interactions were found significant, both genetic effects (between plants. between lines within plants) were consistent from one year to the other, and the ratios of variances appeared rather homogeneous over years. Earlin ess and powdery mildew resistance showed a large epistatic variance. P lant height seemed to be quite additive; this certainly could be relat ed to the presence of two major dwarfing genes polymorphic in the popu lation. Morphological traits of the spike showed larger additive than epistatic variance. Yield components measured on the spike either were predominantly additive or displayed both additive and epistatic effec ts. Far grain yield, which is the most integrative trait, we found lar ger epistatic than additive variance. The genetic control of a given t rait cannot be definitively characterized because it depends on the ge netic material, the test system and the environmental conditions, Howe ver, our results Show the importance of epistasis especially in the ge netic control of complex traits.