Dynamics and persistence in a metacommunity centred on the plant Antirrhinum majus: theoretical predictions and an empirical test

Abstract

Summary Spatial processes have a major influence on the stability of species interaction networks and their resilience to environmental fluctuations. Here, we combine field observations and a dynamic model to understand how spatial processes may affect a network composed of the flowering plant Antirrhinum majus, its cohort of pollinators, and a specialist seed‐predator and its parasitoid. The interactions taking place within this system were investigated by determining the fate of flowers and fruits on flowering and fruiting stems at 16 study sites. We then used this information to estimate spatial and temporal variation in the pollination rate, parasitism rate and hyperparasitism rate. We found that the plants were pollinator‐limited, with relatively variable fruit‐to‐flower ratios across sites. On almost all sites, plants were both parasitized and hyperparasitized, at a low to moderate rate. Comparing our field observations with a tritrophic Nicholson‐Bailey model, we found that empirical data are not always consistent with the conditions for local tritrophic persistence. This suggests that other mechanisms such as random disturbances and recolonizations (patch dynamics) or inter‐site migration through metacommunity dynamics (source‐sink dynamics) play a role in this system. Model simulations showed that dispersal could contribute to increasing tritrophic persistence in this system, and that source‐sink structure, not just environmental stochasticity, may cause the observed pattern of spatial variation. Synthesis. We defined and measured metrics related to species interactions and densities. This led us to suggest that the functioning of the A. majus metacommunity is more consistent with source‐sink than patch metacommunity dynamics, highlighting the extent to which dispersal explains the persistence of the system.