Conclusion
Most studies in invasion ecology aim to find the key drivers of invasion. Such a quest encourages us to focus on main effects and key correlates of invasion, but may poorly equip us to deal with interaction effects and associated context dependence. We contend that flipping our thinking and focusing on mechanisms and factors that limitinvasion, rather than phenomena or factors that drive it, will encourage us to consider how the impacts of various factors vary with a particular context, giving us a stronger foundation for prediction and understanding. As demonstrated in this four-treatment experiment, different invading species experience different levels of seed-, resource- and growth rate-limitation and the hierarchy of limiting factors depends upon the community and sites that the species invade. We found that results largely matched our expectations (Fig. 5). Fast invaders were primarily seed- and resource-limited in early and late successional communities respectively, such that seed dose (propagule pressure) or disturbance would primarily determine their abundance in these two different contexts. Abundance of slow invaders in early and mid succession was limited by invader growth rate such that abundance shortly after introduction may belie long-term trends. As theory would predict, invasion in late succession was primarily resource-limited, such that disturbance that increased soil N strongly facilitated invasion in late succession. Invading plants reduced resident plant diversity via effects on resource limitation, a relationship that was not conditional. Much like in restoration and community ecology, a greater focus on limiting factors in invasion ecology will enable us to better understand why some treatments or hypotheses are important in some situations and not others, and will provide us with a structure for predicting invasion outcomes in different situations.