用主成分和系统聚类法评价棉花品种资源

Assessment of Cotton Variety resources by Principal Component Analysis and Cluster Analysis

本文应用主成分分析法对１０９份引进棉花品种开源的１９个数量性状进行分析，结果前７个主成分对变异的贡献率达８６．７％；从中选择９个综合性状进行系统聚类，１０９份资源被分为二十类，其中Ｃ（尤其ＣⅠ①和ＣⅡ②亚亚类）、Ｅ、Ｌ、Ｐ四类的综合性能较好，且遗传距离较大，在这四类间选亲本杂交成功率较高，且工作量仅为盲目配制组合的８．１７％。本文还就杂交亲本选配的有关问题进行了初步探讨。

Principal component analysis and cluster analysis were performed for 19 traitsof 109 cotton genotypes in this thesis. Results indicated that the contribution of the firstseven principal components to total variation of population studied accounts for over86.7%。The 109 genotypes were grouped into twenty clusters by hierachical clusteringmethod.There were higher probability of selection dominant combination or good cultivarto choose parents for hybridization among different types with good cornprehensive traitssuch as class C,E, L and P. Some problems on parents selection for hybridization werediscussed in this thesis。