作 者 ：申效诚,任应党,王爱萍,张书杰
期 刊 ：生态学报 2010年 30卷 16期 页码：4416~4426
Keywords:geographical distribution, multivariate similarity clustering analysis (MSCA), insect, Henan, China,
摘 要 ：河南省的昆虫、蜘蛛、蜱螨种类共8637种，隶属32目551科3967属。使用多元相似性聚类分析法（MSCA）对全省11个生态区域间的关系进行分析，总相似性系数0.184，表明分布很不均匀，在0.3的相似性水平上，全省分为4个地理分布区：平原盆地区、西北山地丘陵区、伏牛山区和桐柏大别山区。分析了分类阶元、区系成分、不同类群、种类规模、分布地规模对相似性聚类关系的影响，结果显示，属级水平和种级一样，能够揭示省级地理区域内的相似性聚类关系，科级水平将适合更大的地理区域；生物类群是对相似性聚类结构影响最大的参量，保证多类群的参与，是揭示整体规律的关键；在多类群参与的前提下，种类和分布地不是敏感的因素，合适的种类抽样同样能得到满意的结果，分布地的少量增减，也只会引起相似性系数的轻微变动。
Abstract:Henan Province is located at the Central Plains of China. It is not only a transition zone from subtropical to warm regions, but also a transition zone from east plain to west mountanous area. Hence it has a complex ecosystem with high species diversity. Through more than ten years′ investigation, a total of 8637 insect, spider and mite species in Henan Province have been recognized. These species belong to 3967 genera of 551 families in 32 orders, accounting for 11.2% of the total species in China and 0.8% in the world. By using the multivariate similarity coefficient formula and the multivariate similarity clustering analysis (MSCA) proposed by the present authors, all the known species of insects, spides and mites along with a great number of data of the distribution sites were analyzed in Henan Province. The general similarity coefficient of 11 ecotypes in the whole province was 0.184, showing an asymmetric distribution. At the similarity level 0.3, the whole province was divided into four geographic regions, which were Taihang mountainous area in the northwest and the hilly platform area in the west, the north and south slopes of Funiu Mountains, the Tongbo Mountainous area and its adjacent Dabie mountainous area, and the plain and basin region. Clustering the geographically close and topographically similar areas into different distribution regions was very natural and rational, which showed that the analysis method was scientific and the geographical unit and distribution data were applicable. It is superior to the traditional method of clustering analysis, which requires gradual combination and degradation. It should be noticed that, as an accepted boundary between the palearctic and the oriental regions, the Funiu Mountain-Huai River did not give play to the geographical division of Henan as expected. The fact that the south and north slopes of Funiu Mountain was steadily clustered into one region and both banks of Hui River was not divided into two regions showed at least that mountains with an altitude of 2000 m were unable to become distributional barriers of insects, and the narrow Huai River was unlikely to cause biological isolation. To explore the factors affecting the integral similarity, the influence of different taxonomic categories and groups, different fauna elements and scales of distribution sites on the similarity clustering relationships of the whole province was analyzied. The result showed that the increase of the taxonomic categories could increase the similarity and conseal differences; the category at the genus level could effectively reveal the similarity relationships of different ecotypes, while at family level it was not appropriate in analyzing the inter-provincial geographic regions, but could adapt to larger geographic regions. Each fauna has its own distribution patterns, but all except widely spreding species can reveal the similarity clustering relationships that are not very different from the general characters. The biological group is the largest parameter influencing the similarity clustering structure. It is the key factor to open out the macrocosm law by ensuring more groups taking part in. The numbers of species and the distribution sites are not the sensitive factor under the premise that more groups can participant in. Even 10% species can fully post the general clustering characters as long as the increse or decrease of certain species with definite distribution patterns are not involved. So in the similar research, intentionally eliminating the species with definite distribution patterns is unadvisable. Distribution sites are the basic material in calculating similarity. When enough categories and number of species participate in, the slight increase or decrease in the number of distribution sites can only result in a moderate increase or decrease of comparability coefficient, usually will not result in the change of clustering structures. It can be assumed that the basic investigation of the insects fauna and their distribution on both national and provincinal scales has provided enough data to support the similarity analysis and the geographic division. When most groups participate in, it is not necessary to rigidly adhere to include all the species and all the distribution sites, not to say the undescribed species and the uninvestigated sites. The capacity to treat a large quantity of data and the convenience of MSCA method used in this study can provide sufficient technical support to the biogeographic study.
全 文 ：