Use of a Parsimony Algorithm As a Method for Evaluating Ichthyofaunal Distribution and Co-occurrence in an Urban Stream

Abstract

The use of a parsimony algorithm as a method for graphically representing ichthyofaunal distribution and co-occurrence in an urban stream was evaluated using an extensive database of 23 fish species from the freshwater reaches of the Bronx River, New York, USA. The method, based on the idea of considering each sampling station as a "taxon" and the presence or absence of each fish species as a "character state" of that "taxon," resulted in a single most parsimonious tree from which both species distributions throughout the river and the co-occurrence of the fish species at each of eight stations could be clearly visualized and evaluated. The nesting of the stations obtained by this procedure was compared with results from hierarchical clustering, detrended correspondence analysis, and nonmetric multidimensional scaling. This parsimony analysis technique provided information congruent with the above mentioned multivariate methods; in addition, it provided a unique graphical representation of community structure, a representation that is not readily available from traditional multivariate analysis program packages.