Abstract
It has been hypothesized that the relative role of species dispersal in determining plant community composition and species richness changes along primary productivity and disturbance intensity gradients. Manipulative experiments with either diaspore addition or prevention are needed to validate this hypothesis. Due to methodological constraints, diaspore prevention experiments are rarely used. In the case of diaspore addition experiments, there are some potential sources of error. (1) Experiments may be confounded since patchiness of microbial communities is not considered and techniques equalizing microbial communities over study plots are not used. (2) The length of the period of observation may not be sufficient to understand whether the establishment of sown individuals was really successful. (3) The effect of the sowing treatment and the theoretical context of the whole experiment depends on the number and identity of species sown. When addressing the role of long-distance dispersal, it is almost impossible to say what the appropriate number and composition of species to be used for the experiment should be. (4) Until now, most of the attention has been on dispersal in space, while the role of "dispersal in time" (seed bank) has rarely been addressed. We conclude that stepwise accumulation of experimental studies, addressing the role of dispersal in shaping plant communities, will sooner or later reveal general patterns, given the experiments are well planned and aim to avoid the sources of errors described above.