Understanding the mechanisms of species co-occurrence in plant communities and determining the most important drivers of community assembly is one of the central questions in community ecology. Problematics of assembly rules is relatively difficult and most of the studies are based on null models, simulations, or observational methods rather than on experimental approach. This thesis focused on an experimental approach which can clarify many ecological mechanisms contributing to answer many questions related to assembly rules concept. Research in this thesis was focused both on stochastic and deterministic processes influencing the species community composition. Adding seeds or transplants of different species (including also the non-resident species) into community and also into plots with restricted competition and monitoring their establishment and survival for several years, the research highlighted the important role of priority effects on species community composition, the importance of the biotic filter as one of the main drivers in composition of meadow species and suggested that species composition is necessary to compare with the appropriately defined species pool. Research in this thesis also compared different estimation methods of species pool assessment with experimentally determined species pool trying to find the most appropriate possibility of the estimation of species pool.