Abstract: On the basis of comprehensive natural factors including altitude , topography , climate , vegetation , river system , farm belt , etc. , the present paper has divided China's territory into 124 basic units by using the GIS technique. Distribution information of 171 selected mammal species were collected through the China Species Information System (CSIS) , and transformed them into geographical coordinates. Information of presence and absence of species in each unit were used to produce the mammal species distribution similarities to the 124 basic units by using mathematical quantitative analysis , i.e. , S<rensen comparability index formula to calculate correlative matrixes and Ward method for cluster analysis. According to localities of the units that few specimenrecords located , and their similarities to other units , we were able to predict the possible distribution of species. It has found that this prediction method is quite simple , fast , and reasonable accurate.

Key words: Bio-Geographical units, Information System, Mammals similarity cluster analysis, Distribution prediction

摘要： 本研究首先根据综合自然地理因素(包括海拔、地形、气候、植被、水系、农业等) ,利用GIS 技术手段,将中国版图划分成124 个基本单元。选择了171 种哺乳类物种,利用中国物种信息系统收集这些物种的分布信息,并运用GIS 技术将这些信息转换为各个基本单元内这些物种存在与否的信息,再用数学量化分析方法,即S<rensen 多维相似性指数公式计算相关矩阵,以及Ward 方法进行聚类分析,得到这124 个基本单元的哺乳类物种组成相似性聚类图。然后根据少数几个标本采集点所在的基本单元,利用这些基本单元与其他单元的物种组成相似程度,可以预测出物种的可能分布范围。本研究发现,这种预测方法简单,快速而且较准确。

关键词: 生物地理单元, 信息系统, 哺乳动物分布相似性聚类分析, 分布预测