Niche packing and coevolution in competition communities.

Abstract

A coevolutionary model of species packing is developed that allows evolutionary adjustment in both niche position and within-phenotype niche width of one-three competing species. The environment is specified as a single resource dimension x and availability of resources along x is given by a gaussian curve that has parameters x[unk](R) and sigma(R). The model predicts that, for S species, the ratio of optimal niche width w[unk] to sigma(R) is roughly independent of sigma(R) and can be approximated by 1/S when the competitors are completely resource limited. Niche separation (d[unk]/w[unk]) increases only moderately with increases in resource diversity sigma(R) and is greater for two than for three competing species. To the extent that the competitors are not completely resource limited, both coevolutionary niche separation and niche width decrease. Many of the general trends in niche width and niche separation predicted by this coevolutionary model parallel those from optimal foraging theory and limiting similarity models of community structure. The coevolutionary model stands out, however, in the singularly high values predicted for niche separation, making coevolved communities highly invasible. Hence, the theory suggests, as some empirical evidence indicates, that coevolved competition communities can only eixst as such on remote islands or in other habitats that might be free from invasion by outside species.