The study of Carbon Isotope Discrimination under drought stress

Carbon isotope discrimination has been proposed as a criterion for the indirect selection to improve transpiration efficiency and grain yield in bread wheat and barley. Less attention has been devoted to durum wheat (Triticum durum) despite its economic importance in the Mediterranean basin. The ? genetic variation and its relationships to dry matter production and harvest index in durum wheat were investigated in this study. For this purpose, field experiments were conducted on different durum wheat accessions under Mediterranean conditions (South of France) and drought stress. Grain yield, above-ground biomass, harvest index, and carbon isotope discrimination of flag leaf and kernel were measured. Differences between years, noted for both leaf and kernel carbon isotope discrimination were probably related to the variation in water availability from year to year. A large genotypic variation was also noticed for both leaf and kernel carbon isotope discrimination. The two traits were found to be positively correlated with grain yield within and across years, which confirms the interest in carbon isotope discrimination for selection for grain yield improvement under Mediterranean conditions. Both kernel and leaf carbon isotope discrimination correlated better with harvest index than with grain yield, suggesting that carbon isotope discrimination could reflect the efficiency of carbon partitioning to the kernel. The lack of correlation between leaf carbon isotope discrimination and both harvest index and grain yield in favorable water conditions was probably due to the difference in water availability between the period until flag leaves sampling (favorable conditions) and the strong water stress which accompanied the grain filling. Kernel carbon isotope discrimination correlated better with both harvest index and grain yield than did leaf carbon isotope discrimination. Moreover, higher broad sense heritability was obtained for kernel carbon isotope discrimination than for leaf carbon isotope discrimination. As a result, kernel carbon isotope discrimination appeared to be a better predictive criterion for efficiency of the carbon partitioning to the kernel (harvest index), and hence for grain yield, than did flag leaf carbon isotope discrimination.