Interspecific associations and community structure: A local survey and analysis in a grass community

Interspecific associations in the plant community may help to understand the self-organizing assembly and succession of the community. In present study, Pearson correlation, net correlation, Spearman rank correlation, and point correlation were used to detect the interspecific (inter-family) associations of grass species (families) using the sampling data collected in a grass community of Zhuhai, China. We found that most associations between grass species (families) were positive associations. The competition/interference/niche separation between grass species (families) was not significant. A lot of pairs of grass species and families with statistically significant interspecific (inter-family) associations based on four correlation measures were discovered. Cluster trees for grass species/families were obtained by using cluster analysis. Relationship among positive/negative associations, interspecific relationship and community succession/stability/robustness was discussed. I held that species with significant positive or negative associations are generally keystone species in the community. Although both negative and positive associations occur in the community succession, the adaptation and selection will finally result in the successful coexistence of the species with significant positive associations in the climax community. As the advance of community succession, the significant positive associations increase and maximize in climax community, and the significant negative associations increase to a maximum and then decline into climax community. Dominance of significant positive associations in the climax community means the relative stablility and equilibrium of the community. No significant associations usually account for the majority of possible interspecific associations at each phase of community succession. They guarantee the robustness of community. They are candidates of keystone species. Lose of some existing keystone species might be filled with some species previously with no significant associations. In addition, a Java program, associCoeff, re-writed from my earlier work, was introduced. A large number of data were thus given also