Hummon W D
Department of Zoology and Microbiology, Ohio University, 45701, Athens, Ohio.
DAFS Marine Laboratory, Victoria Road, Box 101, AB9 8DB, Aberdeen, Scotland.
Oecologia. 1974 Sep;17(3):203-220. doi: 10.1007/BF00344922.
An index is proposed for measuring similarity between sets of biota over space and time. It integrates into a single measure hitherto separate and often incompatible procedures, based on species presence-absence and species composition. The index is demonstrated by a hypothetical set of data. When compared with the more usual indices of similarity, the proposed index is found to demonstrate greater rigor and hence greater powers of discrimination between data sets.The S index is applied to one series of temporally and two series of spatially differentiated data sets, resulting from sampling the intertidal meiofauna subcommunity of Gastrotricha. All collections represent whole-beach transects, taken at low tide. The temporal series is derived from a boreal-silicious beach in Massachusetts, which was sampled 18 times during the period of a year. The two spatial series are derived from 16 borealsilicious beaches in Massachusetts and New York, and from 8 tropical-carbonate beaches in Florida (6) and the American Virgin Islands (2).Per cent composition by species and H' species diversity values were calculated for data from each site. H' similarity values based on shared species diversity were determined for permutation pairs of sites. These values then were converted into S indices of similarity. The indices were clustered by an unweighted mean, pair-group method and, finally, similarity dendograms were constructed.Results allow the temporal series to be readily separated from both spatial series, and the boreal-silicious spatial series to be clearly distinguished from the tropical-carbonate spatial series. There is strong, though preliminary, evidence that the tropical-carbonate series does not comprise a single biogeographical unit, but that it is subdivided into a Florida subseries and an island subseries. Dendogram analysis indicates that the island effect is nearly as strong as the latitudinal-substratum effect in reducing shared species diversity between intertidal subcommunities of Gastrotricha.
本文提出了一种用于衡量生物群落集合在空间和时间上相似性的指数。该指数将基于物种存在与否和物种组成的、迄今相互独立且通常不兼容的程序整合为一个单一的度量。通过一组假设数据对该指数进行了论证。与更常用的相似性指数相比，发现所提出的指数具有更高的严谨性，因此在区分数据集方面具有更强的辨别力。
S指数应用于一系列时间上和两系列空间上有差异的数据集，这些数据集来自对腹毛动物潮间带小型底栖动物亚群落的采样。所有样本均代表在低潮时采集的全海滩样带。时间序列来自马萨诸塞州的一个北方硅质海滩，该海滩在一年期间采样了18次。两个空间序列分别来自马萨诸塞州和纽约的16个北方硅质海滩，以及佛罗里达州（6个）和美属维尔京群岛（2个）的8个热带碳酸盐海滩。
计算了每个站点数据的物种百分比组成和H'物种多样性值。基于共享物种多样性确定了站点排列对的H'相似性值。然后将这些值转换为S相似性指数。通过非加权平均、成对分组法对指数进行聚类，最后构建相似性树状图。
结果表明，时间序列能够很容易地与两个空间序列区分开来，北方硅质空间序列也能与热带碳酸盐空间序列明显区分。有强有力的（尽管是初步的）证据表明，热带碳酸盐序列并不构成一个单一的生物地理单元，而是被细分为佛罗里达子序列和岛屿子序列。树状图分析表明，在减少腹毛动物潮间带亚群落之间的共享物种多样性方面，岛屿效应几乎与纬度 - 基质效应一样强。
