Studies were conducted with rice varieties developed in recent years in Liaoning Province of China as test materials.Results showed that the varieties tested could be divided into three types: (1) the high-yielding (HYV), (2) the medium-yielding (MYV), and (3) the low-yielding (LYV). The panicle population was in the order of HYV< MYV < LYV, while thegrain number per panicle and seed-setting rate were just opposite. There was no significant difference in the 1 000-grainweight. Biomass was higher in HYV than in MYV and LYV, while the harvest index was higher in HYV and MYV than inLYV. Yield was inversely correlated with panicle population, but positively correlated with the grain number per panicle,seed-setting rate, the number of primary and secondary branch grains, seed density and harvest index. Yield was mostclosely linked with the grain number per panicle which was determined chiefly by the number of secondary branch grainsand to a lesser degree by seed density and the number of primary branch grains. The seed-setting rate of the secondarybranches was positively correlated with that of the whole panicle and yield. The peak value of primary branches for HYV,MYV and LYV was 12-14, 10-13 and 8-12, respectively. LYV had fewer secondary branch grains, while MYV and HYV hadmore. Such grains distributed mainly in the middle and lower parts of panicle axis for LYV and in the middle and upper partsfor MYV and HYV. Studies were conducted with rice varieties developed in recent years in Liaoning Province of China as test materials. Results demonstrated that the said tested products were divided into three types: (1) the high-yielding (HYV), (2) the medium-yielding (MYV), and (3) the low-yielding (LYV). The panicle population was in the order of HYV