结构方程模型在冬小麦农艺性状与产量关系分析中的应用

摘要

为探讨冬小麦主要农艺性状对产量的影响及各性状间的相互作用,采用结构方程模型对2010—2011年度国家冬小麦品种试验中长江上游组(19个品种19个试点)的数据进行了分析,调查性状包括产量(GY)、穗粒数(GNP)、基本苗(BS)、单位面积穗数(SN)、生育期(GD)、千粒重(TGW)和株高(PH).其变异系数为GY>GNP>SN>BS>PH>TGW>GD;与产量的相关程度(相关系数绝对值)为GNP>BS>SN>GD>TGW>PH;在多元回归分析中对产量的效应为SN>GNP>TGW>BS>GD>PH;在结构方程模型中对产量的综合效应(直接效应与间接效应之和)为BS>GNP>TGW>SN>PH>GD.结构方程模型既体现了主要农艺性状对产量的直接效应,也体现了对产量的间接效应,并且作为先验模型,可结合作物生理特性解释主要农艺性状对产量的影响.本研究结果表明,应重视大穗多穗兼顾型冬小麦品种的选育.%This study aimed at understanding the relationship between winter wheat yield and major agronomic traits using struc-tural equation model. The parameters collected from the 2010–2011 National Winter Wheat Region Trail for Upper Yangtze River Group (19 varieties in 19 locations) were grain yield (GY), grain number per spike (GNP), density of basic seedlings (BS), spike number per ha (SN), growth duration (GD), thousand-grain weight (TGW), and plant height (PH). The variance coefficient in structural equation model showed a trend of GY > GNP > SN > BS > PH > TGW > GD. According toPearson correlation, the correlation levels with yield wasGNP > BS > SN > GD > TGW > PH. The effect of a single trait on yield was SN > GNP > TGW > BS > GD > PH according to multiple regression analysis and BS > GNP > TGW > SN > PH > GD according to the sum of direct and indirect effects in structural equation model. Both direct and indirect effects of agronomic traits in winter wheat on yield can be explained by structural equation model. As a prior experimental model, structural equation model can be used to analysis the complex relationship between crop physiological properties and wheat yield. Our results suggest that large- and multi-spikes need to be considered simultaneously in winter wheat breeding.