【摘要】 为提高大规模品种(系)试验的精确度,根据田间土壤肥力连续性原理,建立随机区组设计的一阶、二阶差分近邻小区模型,利用修饰极大似然法估计参数。相对于常规的方差分析,在误差模型与局部控制方面作了改进,从而进一步控制了区组内小区间的异质性。更多还原

【Abstract】 According to the continuity of soil fertility, first difference and second differencemodel of nearest neighbouring plots of randomized block design were proposed for trials with large number of lines in plant breeding, and the parameters of the models were estimated by using the modified maximum likelihood method. In comparison with the classical analysis of variance, the error estimate and the effect of local control were improved, and therefore, the soil heterogeneity among plots in blocks was further controlled.更多还原