Abstract: To study the difference of QTLs for rice yield and its component traits between lowland and upland conditions, a DH (doubled haploid) population including 116 lines, derived from a cross between japonica upland rice IRAT109 and japonica lowland rice Yuefu, was evaluated for plant yield (YP) and its components including 1000-grain weight (GW), seed setting percentage (SP), number of panicles per plant (PN) and grains per panicle (GN). Additive and epistatic QTLs associated with plant yield and its components and interactions between QTL and environment in 2 years were obtained using the constructed molecular linkage map（including 94 RFLP and 71 SSR markers and covering 1535cM. A total of eleven additive QTLs and thirteen pairs of epistatic QTLs under lowland condition and eighteen additive QTLs and seventeen pairs epistatic QTLs under upland condition were detected for plant yield and its components. Two additive QTLs and one pair of epistatic QTLs for 1000-grain weight and an additive QTL for number of panicles per plant were both detected under upland and lowland conditions. Of the eleven pleiotropic or tightly linked QTL regions for yield and its components three regions were also found to be the co-localized regions for root traits. There were eight additive QTLs and eight pairs of epistatic QTLs for plant yield and its components with general contributions of over 10% to phenotypic variation. These QTLs might be useful for upland rice yield molecular marker assisted selection.

Key words: Lowland rice, Upland rice, Yield, QTL mapping, Q×E interaction