VARIABILITY IN SPECIES RICHNESS AND GUILD STRUCTURE IN 2 SPECIES-RICH GRASSLANDS

It has been suggested that variation in the proportion of species in guilds (= guild proportionality) indicates community structuring by guilds in biotic communities. This hypothesis was tested on a subthermophilous grassland and a mesotrophic meadow at a scale of 0.09 m(2) based on a five-year data set. Further, variation in the total number of species, variation in the number of species belonging to a guild and non-randomness in species composition of guilds were studied. A number of criteria for guild definition were used, such as life form, Grime's C-S-R strategy, phenology, plant height, pollination and dispersal syndromes, leaf shape and anatomy and taxonomy at the family level. The observed variation in the number of guild species corresponded to the null model in which species assemblages with fixed species richness per square were randomly generated from the species pool. The observed variation in the number of guild species was often higher than the variation calculated for randomly distributed species whereas the variation in the proportion of guild species was in some cases lower than the variation calculated for randomly distributed species with fixed frequencies. Possible reasons for the discrepancy in the results based on different models are discussed. It is concluded that there is little evidence of guilds in the organization of grasslands.