Competition for benefits can promote the persistence of mutualistic interactions

• We model mutualistic interactions in which individuals compete for benefits.
• We explicitly consider biologically-realistic trade-offs.
• Competition for benefits alone promotes persistence and community assembly.
• Trade-offs can facilitate or constrain persistence depending on their strength.

Mutualistic interactions are characterized by positive density-dependence that should cause interacting species to go extinct when rare. However, data show mutualistic interactions to be common and persistent. Previous theory predicts persistence provided that mutualistic species are regulated by factors external to the mutualistic interaction (e.g., limiting background resources). Empirical data suggest that competition for the benefits provided by mutualistic partners could be a source of negative density-dependence that allows for population regulation, but there is little, if any, theoretical exploration of this mechanism. Here we develop mathematical models to investigate whether competition for benefits alone can allow the persistence of obligate mutualistic interactions. We consider the role of trade-offs in persistence, specifically, trade-offs between benefits acquired versus given and between competition for access to partners (competitive ability) and benefit acquisition. We find that competition for benefits alone is sufficient to promote the persistence of pairwise interactions and the assembly of a three-species community module from an initially pairwise interaction. We find that a trade-off between benefits acquired versus given reduces opportunities for cheating (because a species that acquires significantly more benefits than it gives drives its partner extinct), while a trade-off between competitive ability and benefit acquisition facilitates persistence when it is weak, but constrains persistence when it is strong. When both trade-offs operate simultaneously, persistence requires that each species acquire sufficient benefits to avoid being cheated by its partners, but not so much that it loses its competitive ability. The key finding is that competition for benefits provides a biologically-realistic mechanism for the long-term persistence of mutualistic interactions and the assembly of complex community modules from initially pairwise interactions.