CHANGE IN PATTERN DIVERSITY DURING SECONDARY SUCCESSION IN ESTONIAN FORESTS

Comparisons of observed variance in species diversity (exp H') and in species richness with expectation assuming a random and independent distribution of species are used to assess the relative importance during succession of niche limitation (Wilson, Gitay & Agnew 1987; limitation of the abundance or occurrence of species by competitive interactions) and nucleation (Yarranton & Morrison 1974; the development of a community through chance establishment or persistence followed by vegetative expansion from those nuclei). Wilson, Gitay & Agnew (1987) and Palmer (1987) suggest a deficit (i.e. lower than expected) in variance of richness is consistent with a 'niche limitation' process where competitive sorting makes patches more similar in species number than would be expected. In contrast, variance that is greater than expected is consistent with and could result from underlying environmental heterogeneity (i.e. 'waterhole effect'). We extend these ideas to suggest that, in addition, a deficit in variance of diversity (exp H') can be interpreted as indicating niche limitation. Further, we suggest that a deficit of variance in richness need not be interpreted as resulting from niche limitation, but could result from limited dispersal and establishment followed by localized spread that causes quadrats to have less compositional overlap than might be expected if species were distributed at random. However, there is little reason to expect that such nucleation would cause a similar reduction in the variance of diversity. First, exp H' is relatively insensitive to the presence or absence of rare species, and second, if local competitive processes do restrict the values of exp H' through some form of niche limitation, this could well be independent of the specific species involved. Thus, a combination of lower than expected variance in richness and high variance in diversity suggests nucleation to be important, particularly if the mean pairwise similarity between samples is lower than expected. We first examine a secondary successional sere with stands representing post-logging and mature forest. The variance of diversity is low in the establishment phase, but not in subsequent phases where the tree canopy is well developed. Significantly low variance of species richness is observed in a young forest just entering the thinning phase. Thus, nucleation may be taking place in the establishment phase of forest development, and is almost certainly important in the early thinning phase. Second, we examine communities in the process of equilibration following fertilization, logging, and paludification. All these communities have a low variance in diversity (exp H') suggesting niche limitation.