Grossman argued that communities can be classified as either deterministic or stochastic. Deterministic communities are characterized by persistence (or succession toward a climax) while communities that lack these properties are stochastic. These descriptors were intended to replace the terms equilibrium and nonequilibrium, respectively. Grossman et al.(1982) used this conceptual framework to examine fish collections taken over a 12 yr. period from one stream site in Indiana. They defined community persistence as significant correlation in ranked abundances of 10 seasonally most abundant species, measured with Kendall's coefficient of concordance. Because this test showed no significant similarity among autumn samples, they concluded that this community was not persistent and that 'this assemblage is probably regulated by stochastic factors'. The authors disagree with this conclusion. In choices of study site, sample method, replication, emphasized season, and analytical approach, they believe Grossman et al. took options most likely to show variability. (Copyright (c) 1984 by the University of Chicago).