大豆不同品种(系)性状与产量关系的研究

Relationship Between Characters of Different Soybean Varieties and Yields

本试验以 １９９５～ １９９６年的 １０个大豆品种（系）为试材，利用相关分析、多元线性回归分析、偏相关分析、通径分析等统计方法，对大豆２３个性状进行统计分析，以期找出各性状间及其与产量间的相互关系，为大豆育种及高产栽培提供理论依据。研究结果表明，对单株粒数贡献最大的是主茎二粒荚、三粒荚和分枝二粒荚。对产量贡献最大的是株高、瘪荚、单株粒数、主茎节数和分枝数。欲提高产量时，对植株的高度以直接选择和间接选择兼顾为宜。对株高的直接选择是在有一定株高的基础上，适当降低株高；对株高的间接选择，主要通过结荚高度和百粒重选择。主茎节数以直接增加节数为主，兼顾通过分枝数、单株荚数、单株粒数和瘪荚正向选择主茎节数，分枝数以间接选择为宜，通过主茎节数、单株荚数、株高和瘪美进行选择。对单株粒数应直接选择单株粒数多的材料，对瘪美应直接负向选择。

Fifteen soybean varieties grown in 1995 - 1996 were ed to study the relationships among 23 characters, using statistical analysis methods such as correlation analysis, regression analysis, partial correlation analysis, path -coefficient analysis, cluster analysis. The results showed: 1. Yield as correlated positively and significantly with branches, pods per plant, seeds per plain, effective pods per plants, seeds, one-seed pods, two-seed pods and pods in main stem. 2. The optimal equation of seeds per plant indicated that contributions of two-seed pods and three-seed pods in main stem, two-seed pods in branches were the greatest and their pantal correlation coefficients were significant. 3. The optimal equation of yield indicated that plant height, shrunken pods, seeds per plant, nodes in stem and branches affected greatly the yield and their partial correlation coefficients were significant. 4. The results of path-coefficient analysis between seeds per plant and characters indicated: In order to promote the number of seed per plant, direct positive selection on two-seed pod and three-seed pod in main stem, direct positive and indirect selection on two-seed pod in main branches should be used. 5. In path-coefficient analysis between yield and characters, the influential factors on yield were divided into three aspects: environment, heredity and phenotype. The heredity was the most important one. In order to promote yield, both direct and indirect selection on plant height should be considered. The selection of stem node should be conducted through direct way. Seeds per plant should be selected directly. Shrunken seed should be selected directly and negatively. 6. The results of cluster analysis showed: genetic relationship among varieties used in the test was distant and genetic distance was far.