Abstract
Although many ecologists focus on the relationship between species richness (S) and evenness (E), conflicts between observation and theory are difficult to reconcile. Empirical S-E relationships were not consistent, while relationships show strong correlation between S and E. Since E essentially depended on the relative abundance distribution (RAD), the hypothesis of this paper was that the S-E relationship should be determined by RAD. Theoretical S-E relationships for various RADs have already been reported, but they were rarely assessed by the raw data. This study constructed S-E relationships for a specific RAD, which indicated that if the community had a fractal distribution of rank abundance, E would decrease with S, and the S-E relationship would be unique for a given RAD. Such theoretical expectations were supported by three datasets with 82 samples, which suggested that the S-E relationship were controlled by RAD and inconsistent S-E relationships in statistical analyses could be accounted for by the variation of underlying RAD model between communities. From the perspective of RAD, it could be too early to split the diversity into S and E only based on the S-E relationship in statistical analyses.