昆虫区系多元相似性分析方法  

A discussion about the method for multivariate similarity analysis of fauna

作　　者：申效诚[1] 孙浩[1] 赵华东[1] 

机构地区：[1]郑州大学生物工程系

出　　处：《生态学报》2008年第2期849-854,共6页Acta Ecologica Sinica

基　　金：河南省自然科学基金项目(411031700)~~

摘　　要：由植物学领域首先提出的相似性概念已广泛应用于动植物区系地理研究以及生物学、生态学等诸多自然学科乃至社会科学领域。根据Jacard提出的二元相似性系数公式SI=C/(A+B-C)和Sprensen提出的二元相似性系数公式SI=2C/(A+B),分别推导出2个计算多元相似性系数的数学表达式,SIJab…n=[(∑Hij)2/n+(∑Hijk)3/n+…+Hab…n]/[∑Si-∑Hij-2∑Hijk-…-(n-1)Hab…n]和SISab…n=[2(∑Hij)2/n+(3∑Hijk)3/n+…+nHab…n]/∑Si,并用中国夜蛾广布种类在中国7个动物地理区的分布资料为例进行验证,从而可以直接从整体角度和宏观规模上简便、快捷地考量多个系统间的亲疏程度和相似关系。建议在以相似性为基础的聚类分析中,不必再先把2个系统合并成一个新系统后,再和第3个系统比较,而可直接计算多个系统的相似性系数,以避免由于合并带来的信息损失。还讨论了应该提高Sprensen公式0.5的显著性标准,以使同一组数据的两种计算结果趋向一致。The concept of similarity, originated from studies in botany, but has been widely applied in plant and animal geology, biology, ecology and even sociology. Based on the two bivariate analysis formulas SI = C/（A ＋ B- C） and SI =2C/ （ A ＋ B ） , two expression equations ,SIJab…n=[（∑Hij）2/n＋（∑Hijk）3/n＋…＋Hab…n]/[∑Si-∑Hij-2∑Hijk-…-（n-1）Hab…n]和SISab…n=[2（∑Hij）2/n＋（3∑Hijk）3/n＋…＋nHab…n]/∑Si were developed to describe the coefficient of similarity for muhivariate data. These two equations were validated using population distribution data of the common Noctuidae species from seven animal geological areas in China. By using these two equations, the species similarity and the relationship distance in the whole system overall can be efficiently measured. We suggest how to appropriately use these two equations to enhance the significant standard to 0. 5 in Spensen＇s formula, so that a consistent result for the same set of data can be obtained from both formulas.

关 键 词：相似性 多元 公式 昆虫区系 

分 类 号：Q968[生物学—昆虫学]