Abstract: To facilitate breeding for high grain yield potential in rice, two BC3F5 introgression populations, derived from two crosses between Fengaizhan (FAZ, the recurrent parent) and two donors (Khazar and IR64), were genotyped with 89 and 101 polymorphic SSR markers and evaluated in two environments (early and late seasons growing in Guangzhou) to identify and map QTL affecting grain yield (GY) and its components, panicle number per plant (PNP), filled grain number per panicle (GNP), and 1 000-grain weight (TGW). Regression analyses indicated that GY was largely determined by GNP followed by PNP in the early season, and by PNP in the late season. The contribution of TGW to GY was small in both seasons. ANOVA analyses revealed a total of 27 QTL regions associated with GY and its components across the rice genome, including 10 GY QTL, 9 PNP QTL, 9 GNP QTL, and 14 TGW QTL. Fifteen (35.7%) of the identified QTL mapped to the similar regions with previously reported QTL affecting the same traits, indicated the robustness of our results. The majority of the QTL were only detected in one of the envi-ronments, indicating a strong component of G × E interactions of the identified QTL. Three interesting genomic regions on chro-mosomes 3, 7, and 9 were noted, which were associated with GY and two of its components. Detailed analyses indicated that both pleiotropy and linkage might be responsible for the observed associations. Our results provided useful information to our current understanding of the genetic basis of GY in rice, and a new strategy of using IL populations for both QTL discovery and breeding.

Key words: Rice, Introgression lines, QTL, Grain yield and components, G ´, E interaction, Pleiotropy