Convergent high diversity in naturally colonized experimental grasslands is not related to increased productivity

•Initial plant diversity affects subsequent community assembly processes.•Species richness converges at high levels due to the accumulation of colonists.•Species accumulation is paralleled by increased taxonomic and phylogenetic diversity.•Trait diversity is limited by competitive exclusion at high initial plant diversity.•Convergent high levels of diversity do not result in a maximization of productivity. Initial plant diversity might control subsequent community assembly processes and plant productivity. To study these effects, we used a biodiversity experiment (Jena Experiment) with subplots of different sown diversity that were never weeded and spontaneously colonized control plots of different size (3.5×3.5m, 20×20m) with and without mowing in an 8-year study. On non-sown bare plots without mowing, colonizer accumulation depended on plot size resulting in a loss of diversity in large, undisturbed (unmown) control plots after initial colonization. On sown plots that were mown, species richness converged to high levels due to the accumulation of internal colonists (species belonging to the experimental pool of sown species), while initially high species richness of external colonists (species not belonging to the experimental pool) and residents (species sown on the plot) declined over time. The convergence of total species richness at higher levels was paralleled by increased taxonomic (Simpson index, QSimp), phylogenetic (QPhylo) and trait (FDQ) diversity, whereby FDQ was greatest on plots with low resident species richness after several years. Rates of change in terms of species colonization and extinction decelerated over time irrespective of resident species richness, mowing or plot size. While the contribution of residents declined and that of colonists increased, community biomass production did not change over time. The biomasses of residents and colonists were greatest at higher levels of species richness, partial QPhylo and FDQ of the respective species group, but community biomass was consistently weakly related to total species richness and diversity indices. Our study shows that in contrast to the period of succession, “mature” plant communities resulting from natural assembly processes, favouring the coexistence of multiple species and thus high biodiversity, do no longer show significant relationships between species richness and variables related to ecosystem functioning such as primary productivity.