Several experiments demonstrate that the functional characteristics of invading species relative to those already in a community can help determine initial success of those invaders. Less clear, however, is if persistence of those invaders follows similar patterns over the long-term. We assessed the success of six focal species of plants seeded into different serpentine grassland communities, and followed their abundances during 6 years of natural community assembly. Because this is predominantly an annual system, responses involved both growth of individuals and realistic population dynamics over several generations.
Results/Conclusions

Initial success of the invaders was predicted well by the functional composition of the original community. However, we observed different patterns among the group of focal species over the long term. Some invaders (e.g., Centaurea solstitialis) declined to very low abundances in all communities as community assembly progressed and overall plant densities increased. In such cases, initial community composition no longer predicted abundance after a couple of years. For others (e.g., Bromus madritensis ssp. rubens), abundances generally increased and continued to follow patterns caused by initial community composition for up to four years. By the sixth year of the experiment, original community composition still had significant effects on invader abundance, but in a different pattern from that seen initially. While success of the B. madritensis (an early-season annual) initially was lowest in communities already containing other early-season annuals, by the sixth year, such treatments had the highest abundances of B. madritensis – in part because of declines in the abundance of this species in other treatments. These results suggest that 1) initial community composition can have strong effects on community assembly, but 2) long-term patterns may not follow initial patterns.