Seed dispersal by ants is a key ecosystem function in North American eastern deciduous forests, where 30-40% of understory forbs are “myrmecochores” or adapted to ant-mediated seed dispersal. In this region, Aphaenogaster sp. are responsible for the majority of seed dispersal and are considered “keystone dispersers.” While this mutualism is relatively well-studied, we have limited understanding of how it functions across its range, including along variation in anthropogenic disturbance. In the northeast, most deciduous forests are “secondary” or previously cleared for agriculture and allowed to passively regenerate. Understory plant communities are not resilient to this disturbance with low recovery, especially of myrmecochore species. While several factors lead to low recovery, a reduction in this function is a likely impediment. I am investigating if ant-mediated seed dispersal is intact in secondary forests. To this end, I have conducted a large-scale natural experiment in 20 primary and secondary forest sites in the northeast. I conducted seed dispersal trials, sampled the ant community using pitfall traps, measured the presence and cover of understory plant species, and quantified ant habitat. I am examining if several factors, including land use history, affects dispersal rates, the diversity and abundance of myrmecochores, and their mutualist ant partners.
Results/Conclusions
I performed generalized linear models (GLMs) on the proportion of seeds dispersed by ants, and the abundance of Aphaenogaster sp. with forest land use history type (secondary vs. primary) as a fixed effect and region as a random effect. I am finding higher and more consistent seed removal in primary forests and lower and more variable removal in secondary forests (p < 0.001). Aphaenogaster sp. abundance was higher in primary forests than in secondary forests (p = 0.043). For both results, models including region were better fits (using AIC), meaning that region also explains variation in responses. To determine what factors influenced the observed variance in seed removal, I conducted a path analysis to test if direct and indirect effects between the abundance of seed mutualists and antagonists and habitat factors affects removal rates. My final path model revealed that Aphaenogaster sp. abundance had a direct positive effect on removal ( p < 0.01) and that seed antagonists have a weak negative impact on removal. Finally, I conducted a Principal component analysis (PCA) on cover of myrmecochore species, and I found differences in PC1 between forest types (p = 0.007). Future analyses will explore factors that affect myrmecochore diversity.