杏树不同品种果核数量性状分析

以20个杏树品种为试验材料，运用SPSS19.0软件，分析研究了杏树果核数量性状指标，对测得果实的核纵径、核横径、核侧径、核型指数、核干重、仁纵径、仁横经、仁侧径、仁型指数、仁干重以及出仁率11项果核性状指标进行了主成分分析和聚类分析。结果表明，前3个主成分累积方差贡献率可达84.55%，囊括了所检测的11个果核性状指标。对20个杏树品种，使用前3个主成分的特征值进行系统聚类分析，欧式距离为10时，该20个杏树品种可聚为5类。对11个果核性状指标，利用前3个主成分的特征值进行系统聚类分析，在相同欧式距离处，该11个果核性状指标可聚为4类，分别从这4类指标中挑选仁侧径、核侧径、出仁率、仁型指数4个具有代表性的指标，对20个杏树品种进行系统聚类, 在欧式距离为10时, 聚类结果与用3个主成分值聚类结果相同，表明用以描述杏树果核性状的11个指标可简化为4个指标，即仁侧径、核侧径、出仁率、仁型指数。为其他经济林树种的性状指标筛选简化提供参考依据。

Quantitative Characters of Apricot Kernels from Different Cultivars

In order to simplify the characteristic indexes of fruit without decreasing the information volume contained in original indices, multivariate statistics methods of principal component analysis and systematic cluster analysis were used for analyzing 11 characteristic indexes of fruit, including nut vertical diameter，nut crosswise diameter，nut diameter side，nut shape index，nut dry weight，kernel vertical diameter，kernel crosswise diameter，kernel diameter side，kernel shape index，kernel dry weight and percentage of kernel for 20 apricot varieties. The results indicated that the cumulative contribution of the first 3 principal components reached 84.55% and involved in 11 indices. Systematic clustering analysis using the 3 components as variants，when the Euclidean distance approached to 10，the 20 apricot varieties could be divided into five categories. Principal component cluster analysis of the first 3 component variants showed that the 11 fruit character indices could be clustered into 4 types at the Euclidean distance of 10．And the 11 fruit characteristic indexes could be simplified as kernel vertical diameter，nut diameter side，percentage of kernel and kernel shape index. However, using the 4 indices as variants in systematic clustering analysis on 20 apricot varieties，when the Euclidean distance was around 10, the clustering results resembled to the one by using the 3 component as variants with all 11 indices included, which suggested that the 11 indices might be simplified and represented by the 4 ones (kernel vertical diameter, nut diameter side，percentage of kernel and kernel shape index). The results could provide a possibility to simplify the character indices for others crops.