Interpreting yield instability in pea using genotypic and environmental covariates

The improvement of yield stability in many grain legume species and especially in dry pea (Pisum sativum L.) is a major breeding objective. It is therefore essential that breeders understand the basis of the yield instability phenomenon in order to develop stable and high yielding cultivars. The study of 10 genotypes, sown at two dates (early/late), during 2 yr (1991-1992), in France, at three I.N.R.A. experimental stations (Mom, Versailles, Dijon), confirmed the existence of high variability for yield. The environmental effect was preponderant and essentially due to differences of soil water availability during the flowering period. Genotype x environment (g x e) interaction also played an important part in yield variability. A factorial regression analysis was conducted with genotypic and environmental covariates to determine the genotypic morphological traits and the environmental factors involved in the g x e interaction. Three covariates played an important role: node number of the first flower (N1F) and the mean number of reproductive nodes (NR) of the main stem, as genotypic covariates, and the meteorological water balance during the flowering period (MWBF), as an environmental covariate. These variables explained 64% of the total g x e interaction. NIF and NR define, respectively, the beginning and the duration of the seed set period on the main stem of the plant. These results suggest that an important part of the g x e interaction in pea is determined by the differential response of genotypes, according to their earliness to flower (estimated by NIF) and the duration of the seed set period (determined by NR), when grown in environments with drought stress during the sensitive flowering period.