文献类型：期刊文献

中文题名：薄壳山核桃无性系果实性状指标简化研究

英文题名：A Study on Simplifying the Indices of the Clone Fruit Characters of Carya illinoemis

作者：李川[1,2] 姚小华[1] 王开良[1] 辜夕容[2]

第一作者：李川

机构：[1]中国林业科学研究院亚热带林业研究所;[2]西南大学资源环境学院

年份：2011

卷号：33

期号：4

起止页码：696-700

中文期刊名：江西农业大学学报

外文期刊名：Acta Agriculturae Universitatis Jiangxiensis

收录：CSTPCD;;北大核心:【北大核心2008】；CSCD:【CSCD_E2011_2012】；

基金："948"项目(2006-4-82);农业科技成果转化资金项目(2006GB24320401)

语种：中文

中文关键词：薄壳山核桃;无性系;主成分分析;聚类分析

外文关键词：Carya illinoem is; clone; principal component analysis; clustering analysis

分类号：S664.1;Q949.735

摘要：利用SPSS软件对17个薄壳山核桃无性系10个果实性状指标(果重、果高、果径、果型指数、果皮厚、核重、核高、核径、核型指数以及出核率)进行了主成分分析、聚类分析。结果表明,前3个主成分就能概括所有性状98.47%的信息。利用这3个主成分值对17个薄壳山核桃无性系进行系统聚类,在欧式距离为10时,无性系12号、17号、6号、16号、3号、1号、5号、4号、13号、7号和14号为一类;无性系8号和9号为第2类;无性系2号、11号和15号为第3类;无性系10号单独为第4类。对前3个主成分的特征值进行果实性状的聚类,在欧式距离为10时,可以将10个果实性状指标聚为4类,从这4类果实性状指标中选出果高、果重、核重和出核率等4个指标对17个薄壳山核桃无性系进行系统聚类,在相同的欧式距离,聚类结果与3个主成分值的聚类结果相符。因此,可以将薄壳山核桃无性系的10个果实性状指标简化为4个性状指标,即果高、果重、核重和出核率,同时也为其他作物性状指标简化提供了一个可能。
Principal component analysis and clustering analysis were applied to analyze 10 fruit character indices（fruit weight,fruit height,fruit diameter,fruit shape index,peel thickness,nut weight,nut height,nut diameter,nut shape index,nut weight in fruit） of 17 Carya illinoemis.The results indicated that,the first 3 principal components can explain 98.47% of 10 characters.Systematic clustering analysis using the 3 components as variants,when the euclidean distance was around 10,clone fruit no.12,17,16,3,1,5,4,13,7 and 14 were classified as one group,and clone fruit no.8 and 9 as one group,clone fruit no.2,11 and 15 as one group,and clone fruit no.10 as one group.Principal component cluster analysis of the first 3 component variants showed that the 10 fruit character indices can be clustered into 4 types at the euclidean distance of 10.Four indices,namely,fruit weight,fruit height,nut weight and nut weight in fruit,chosen from the four types of fruit characters,were used as variants in systematic clustering analysis,at the same euclidean distance,the result was the same as the clustering results with 3PC as variants.So the 10 fruit character indices can be