摘要
应用数值分类的多种方法(系统分类法、模糊分类法、图论分类法、主分量分类法)对我国150种槭树进行了数值分类研究。根据各种数值分类方法得到的结果,以协表相相关系数作为最佳分类的依据,确定了相关系数UPGMA法的分类结果为最佳。画出树系图,找出属以下分类阶元的结合线。根据结合线,并考虑该属进化系统学知识,最后将槭属分为14个组。与传统分类系统的分类结果相比较,绝大多数种相符,并且使有的类群得到合理的归并和划分。同时,对近年发表的近几个新种也进行了归类。
150 Species of Chinese maples have been studied by means of numerical taxonomic methods in this paper. The methads used here are cluster analysis, fuzzy cluster analysis, minimal spanning tree and principle component analysis. The characters for classification were mainly selected from external morphology. After standardization of the characters of the original data matrix, we computed the correlation and distance coefficients various methods of classification were used Q classification was performed in turn on correlation and distance matrix. The cophenetic correlation coefficient were used to get the optimal classification and the comparison of cluster methods showed that the unweighted pair-group method using arithmetic averages (UPGMA) on the correlation matrix may be considered as the most suitable one for Q cluster in biology. The results indicated that the 150 species of the Chinese Maples may be divided into 14 groups. This is on the whole in agreement with the traditional Classification. with the help of Dendrongram, it will be quite clear at a glance the interrelationship of the genus Acer classification of species or groups are reasonably arranged in order which could not be achieved by traditional classification. In comparison, the fuzzy and the graph theoretical methods of classifiation are also usable.
