| Abstract: | The data obtained during an extensive multidisciplinary investigation of a polluted harbor basin in Hamburg were analyzed in various ways in order to determine which statistical methods are easiest to interpret and permit the most accurate evaluations. The methods analyzed include the species deficit, index of species similarity, and saprobity index, as well as the more complex cluster and gradient analyses, which revealed a variety of relationships among the biotic communities at the various sampling sites and some ecological tolerance limits of various species. The shortcomings of numerical indices for characterizing water bodies are discussed. It is suggested that biostatistical methods are much more valuable for the empirical evaluation of data than for artificially classifying ecosystems. |
