What do the biodiversity experiments tell us about consequences of plant species loss in the real world?

Experiments where the diversity of species assemblage is manipulated are sometimes used to predict the consequences of species loss from real communities. However, their design corresponds to a random selection of the lost species. There are three main factors that limit species richness: harshness of the environment, competitive exclusion, and species pool, limitation. Species loss is usually caused by increasing effects of these factors. In the first two cases, the species that are excluded are highly non-random subsets of the potential. species set, and consequently, the predictions based on random selection of the lost species might be misleading. The data show that the least productive species are those being recently excluded from temperate grasslands and consequently, species loss is not connected with decline of productivity. The concurrent species loss in many communities, however, means also a reduction of the available diaspore pool, on a landscape scale, and could result in increased species pool limitation in other communities. (C) 2004 Elsevier GmbH. All rights reserved.