摘要
以67份四川小麦地方品种为试验材料进行农艺性状分析,并对性状之间进行简单相关和偏相关分析,多元逐步回归和通径分析,旨在为小麦的遗传改良与育种利用提供依据。结果表明:四川小麦地方品种植株总体偏高,有效穗数多数在10个以下,小穗数平均为20.4,部分品种穗粒数较多、大于70粒,但千粒重和单株产量较低,平均值分别为27.5和11.2 g。筛选出一些单一或综合农艺性状优良的材料。性状相关分析表明:随着有效穗数、穗粒数和千粒重增加,单株产量显著提高。逐步回归分析和通径分析都表明:有效穗数、穗粒数和千粒重对单株产量的正向直接影响较大。多元回归方程:Y=-21.96+1.54X2+0.18X3+0.42X5解释了单株产量变异的97.09%。因此,高产育种时,宜考虑选择有效穗数、穗粒数和千粒重高的小麦品种。
The agronomic characters of 67 Sichuan wheat landraees were evaluated by using the statistical methods of simple and partial earrelation analysis, multiple stepwise regxession and path analysis in order to provide helpful information for the genetic improvement of wheat. The results indicated that the plant height of most landraces was too high to be used for breeding progam, and the spike number per plant of most landraces was less that ten.The average number of spikdets was 20.4. The grain number per spike of some materials was more than 70,but the 1000-grain weight and gain weight per plant of them were lower, with the mean of 27.5 g and 11.2 g,respectively.A few landraees were selected out based on their single or general agronomic characters. Correlation analysis of characters indicated that with the increase of spike number per plant, grain number per spike and 1000-grain weight, the grain weight per plant would increased obviously. The stepwise regression and path analysis of agronomic charactem showed that the positive effects of spike number per plant, grain number per spike and 1000grain weight on grain weight per plant were larger. And 97.09 % variation of grain weight per plant could be explained by the multiple regression equations Y = - 21.96 + 1.54X2 + 0.18X3 + 0.42X5 Therefore,the varieties with higher spike number per plant, higher grain number per spike or higher 1000-grain weight could be considered to be utilized in high yield breeding.
出处
《西南农业学报》
CSCD
2006年第5期791-795,共5页
Southwest China Journal of Agricultural Sciences
基金
高等学校博士学科点专项科研基金(20040626004)
四川省教育厅科研项目(2005B003)资助
关键词
小麦
农艺性状
相关分析
通径分析
回归分析
wheat
agronomic characters
correlation analysis
path analysis
regression analysis