An attempt is made to derive a measure for the degree of plant community organization, which would not be based on species co-occurrence and co-abundance. The use of the independent random distribution hypothesis (IRDH) is suggested for this purpose. The hypothesis is expected to be valid, if no deterministic phytosociological-structure-generating mechanism is present. If structural variability is used as a statistic for testing the hypothesis, deviances from the conditions of IRDH (species distributions are independent from each other, environmental gradients are lacking) will be attributable either to species interactions (smaller structural variability than expected), or to environmental heterogeneity (greater structural variability than expected). Structural variability is evaluated as the variance of species diversity, the index N = exp(H') is used for measuring diversity. The precise measure of the degree of community organization Wis computed as the shift between two empirical distributions: D* (VN) or Bootstrap distribution of variance of diversity in the community, and DΒ°(VN) or 'the random community variability' distribution, which is evaluated after simulating the IRDH conditions.
A satisfactory interpretation can be given to the results of evaluating W for 11 data sets of 10 relev6s each.