The co-distribution of species in relation to the neutral theory of community ecology

Patterns of species co-distribution have often been used to infer the operation of processes such as competition and selection among species. Ecological specialization should create unexpectedly strong species associations, reflected by high positive or negative correlations between species that are adapted to similar or different kinds of site. Two systematic surveys of plant communities do indeed show stronger associations than are expected in randomly assembled communities. Neutral community models, however, also predict nonrandom patterns of co-distribution. Local selection in heterogeneous landscapes, where species are sorted into the sites to which they are best adapted, will produce even stronger associations, provided that different kinds of site are correlated in space, forming repeated combinations or "habitats." In practice, however, this effect is readily detectable only for intense selection in coarse-grained landscapes. In contrast, classification of species or sites into prior categories consistently produces positive associations, and thereby demonstrates the existence of non-neutral processes structuring communities. It is concluded that the bulk properties of communities, such as overall diversity, may be strongly affected by local dispersal and stochastic drift and, for this reason, are adequately represented by neutral models in many cases, despite the operation of systematic processes of local selection. One corollary of this interpretation is that these bulk properties are unlikely to provide useful information about community processes.