Plant richness and composition within a native functional group control invasion success in grassland microcosms

Background/Question/Methods

Plant community diversity can provide resistance against exotic species invasions, although the relative roles of species richness and composition are still unclear. Certain functional groups may be particularly important to control invasion magnitude, and therefore loss of diversity within such groups would be critical. In the Flooding Pampa of eastern Argentina, grassland invasibility by annual exotic species has been found to be negatively correlated with native warm-season (WS) grass richness across sites on a regional scale. Here we test the hypothesis that invasion magnitude is affected by both the richness and specific composition of resident WS grasses. We established 17 plant mixtures with one (n=9), three (n=4), or six (n=4) grass species, thus creating low, medium, and high-diversity microcosms, and a bare-ground control with no natives (n=6). The medium and high-diversity treatments were each represented by four species mixtures drawn randomly from a pool of nine summer grasses. All species were also grown in monocultures. Each mixture was replicated three times, for a total of 57 microcosms. Plant mixtures were established in 10-L pots with the same initial density of grass ramets. Microcosms were sowed with the annual Lolium multiflorum, a widely invasive exotic grass in the pampas.

Results/Conclusions

Presence of native grasses reduced seedling size, total shoot biomass, and fecundity of L. multiflorum by 40–60% (all P<0.0001), but did not alter invader seedling recruitment (P>0.10), relative to bare-ground controls. Biomass of L. multiflorum significantly decreased with native grass richness (P<0.05), yet it was more strongly affected by species composition (P<0.0001). Invader fecundity depended on species composition (P<0.02) and not richness (P<0.09). The least invaded medium and high-diversity mixtures supported significantly less L. multiflorum biomass and fecundity than the most invasion resistant monoculture, that of Stenotaphrum secundatum. However, when presence of this prostrate grass was added as a predictor of invader biomass (P<0.0005), richness was no longer significant (P=0.43), while the overall effect of composition was only slightly modified (P<0.003). These results indicate that a “selection effect” operated across the richness gradient, which did not account for all observed biodiversity effects on invasion. Diversity effects became apparent only after winter, when WS grasses begin to growth actively. Thus, resource competition, rather than recruitment inhibition, limited invader performance in certain species-rich mixtures. We conclude that both resident grass richness and composition control invasion by L. multiflorum. Yet, understanding mechanisms for differential biotic resistance among grass assemblages would require trait-based functional analyses.