Interaction networks among species may develop over time in plant communities, making them more productive and stable. Species complementarity increases over time as shown in several biodiversity–ecosystem functioning experiments. This increase has been interpreted as a potential consequence of complementary resource use or pathogen load among species. It could, however, be due to selection for increased “combining ability” in populations initially containing members that perform well in monocultures or low-diversity mixtures, or in mixtures. Those population members selected for would become dominant. To test this hypothesis, we collected plants from 8-yr old experimental grassland communities of varying species richness and composition established in 2002 at a field site in Jena, Germany. We grew the plants in monoculture or two-species mixture in a glasshouse, in a constant pot density of four individuals. The design included twelve species representing four functional groups; all possible combinations of the functional groups were attempted in the mixtures. The data were analyzed with mixed-effects models, using monoculture vs. mixture community history and planted diversity and their interaction as fixed terms; species identities and mixture combinations as random terms. An increased combining ability would be indicated by a significant interaction between community history and planted diversity.
Results/Conclusions
Pots containing two-species mixtures produced more aboveground biomass than those containing monocultures. However, as predicted, this positive effect of planted diversity on community productivity was stronger for plants with a history of growth in mixed communities than for plants with a history of growth in monocultures (significant community history x planted diversity interaction). This result indicates that species complementarity and combining ability decreased in monocultures and increased in mixed communities over the 8-year time span of the Jena experiment. Combining ability was increased to the same extent in species mixtures containing a single or several plant functional groups. Under the assumption that the original seed material used to establish the field experiment included some genotypic diversity for each of the twelve tested species, we hypothesize that genotypes that can grow well in monocultures became more abundant and were therefore more likely to be collected in monocultures and vice versa for mixtures. We conclude that increasing complementarity effects over time in biodiversity experiments may at least in part be due to the evolution of increased combining ability.