Historical forest disturbance results in variation in functional resilience of seed dispersal mutualisms.

Mutualistic interactions provide essential ecosystem functions that contribute to promoting and maintaining diversity in ecosystems. Understanding if functionally important mutualisms are "resilient" (i.e., able to resist or recover) to anthropogenic disturbance is essential for revealing the capacity for diversity to recover. Animal‐mediated seed dispersal supports plant population growth and influences community structure, and disturbance affecting seed dispersal can contribute to low resiliency of plant diversity. Ant–seed dispersal mutualisms are sensitive to anthropogenic disturbance, as they rely on one to a few high‐quality dispersal partners. In North American eastern deciduous forests, ants in the genus Aphaenogaster are "keystone dispersers" of understory forbs adapted to dispersal by ants (myrmecochores), which make up more than one‐third of the understory herbaceous community. The majority of forests within this region have regenerated from previous disturbance in the form of clearing for agriculture. Previous studies have revealed that myrmecochore diversity is not resilient to previous clearing. Here, we ask if seed dispersal mutualisms are resilient to historical forest disturbance and if decreases in mutualistic interactions with partners, Aphaenogaster sp., or increases in antagonistic interactions cause degradation of function. In a large‐scale natural experiment (20 sites), we measured seed removal, the abundance of mutualistic partners and other invertebrates interacting with seeds, myrmecochore cover, and diversity, along with ant habitat and forest structure. We found lower and more variable seed removal in secondary forests compared with remnant forests. A path analysis of all forests revealed that the abundance of mutualists was the primary determinant of the variation in seed removal, and that seed damage by antagonists (invasive slugs) negatively affected dispersal and was higher in secondary forests. In a path analysis of remnant forests, the link between mutualist abundance and seed removal was absent, but present in the secondary forest path, suggesting that seed dispersal is more variable and dependent on the mutualist abundance in secondary forests and is stable and high in remnant forests. Our results suggest that functional resilience to disturbance is variable, where seed dispersal is low in some secondary forests and not others. This work provides key insights into the effects of disturbance on mutualistic interactions and how the resilience of critical ecosystem functions impacts the capacity for diversity resiliency. [ABSTRACT FROM AUTHOR]