SPECIES-POOL HYPOTHESIS: LIMITS TO ITS TESTING

The species richness of a community depends both on the pool of available species and on biotic mechanisms that lead to the exclusion of some of the species from a community. The method suggested by PARTEL et al., Oikos 75: 111-117, 1996 to test the effect of species-pool size on the species richness of a community is discussed in this paper. This method is based on the calculation of a correlation between the actual species richness of a community and the actual species-pool size, and compares the correlation found in the data with that simulated by a null model. In the null model, the species richness has a uniform distribution between zero and the size of the species pool. A correlation significantly higher than that in the null model is interpreted as evidence of the greater role of species pool than of local interactions in formation of community species richness (ZOBEL, Folia Geobot. 36: 3-8, 2001). It is shown that the interpretation of discrepancies between the null model and reality is difficult or impossible, because: (1) a null model with a uniform distribution of species richness is unrealistic, (2) both models based on the random selection of species from a species pool and models that include competitive interactions in the community predict a higher positive correlation of species richness and size of the species pool than the null model, and (3) local species richness might be affected by species-pool size, but a large species-pool size can also be a result of high local species richness. Caution is urged when interpreting the analyses based on the size of the filtered species-pool size.