ANALYSIS OF QTL X ENVIRONMENT INTERACTION FOR YIELD COMPONENTS AND PLANT HEIGHT IN RICE

An F-2 and two equivalent F-3 populations of an indica-indica cross of rice, Tesanai 2/CB, were constructed and grown in different environme nts. The identification of quantitative trait loci (QTL) for yield com ponents and plant height and an analysis of QTL x environment interact ion were conducted for three trials. Interval mapping of QTL for eight traits was employed with a threshold of LOD = 2 using the computer pa ckage MAPMAKER/QTL. A total of 44 QTL were detected in 18 intervals of nine chromosomes, including 3 for the number of panicles (NP), 5 for the number of filled grains (NFG), 6 for total number of spikelets (TN S), 3 for spikelet fertility (SF), 7 for 1000-grain weight (TGWT), 5 f or grain weight per plant (GWT), 8 for plant height (PH) and 7 for pan icle length (PL). The numbers of QTL detected in two or three trials w ere 1 for NP, 1 for NFG, 1 for TNS, none for SF, 4 for TGWT, 3 for GWT , 2 for PH and 5 for FL, making a total of 17. When a QTL was detected in more than one trial the direction and magnitude of its additive ef fect, the dominance effect and the degree of dominance were generally in good agreement. In all three trials, QTL were frequently detected f or related traits in the same intervals. The directions of additive ef fect of QTL for related traits in a given interval were in agreement w ith few exceptions, no matter whether they were detected in the same t rial or not. This result suggested that pleiotropism rather than close linkage of different QTL was the major reason why QTL for different t raits were frequently detected in the same intervals. When gene pleiot ropism was considered, 23 of the 29 QTL for yield and its components a nd 9 of the 15 QTL for plant stature were detected in more than one tr ial. This indicated that the detection of chromosomal segments harbori ng QTL was hardly affected by environmental factors.