The Janzen-Connell hypothesis is one of the most important hypotheses in ecology, but is mostly tested indirectly without accounting for the underlying plant-associated organisms, or only for highly host-specific organisms. Advances in massive sequencing allow to sample, for example, the fungi communities associated with different plant species, and spatial analysis can reveal spatial patterns in the number of organisms (species) shared by plant species and their neighbours. We show how combining these tools provides new perspectives for testing the Janzen Connell hypothesis. We illustrate our approach using a natural experiment in two fully-mapped Mediterranean forest plots, where the dominant dry- and fleshy-fruited species have distinctly different seed deposition patterns, leading to contrasting expectation about the emerging spatial structures. Our analysis confirmed these expectations and provided deep insights into how the neighbourhood load of plant-associated fungi and herbivorous insects changes during plant ontogeny and how seed dispersal mechanisms modulates Janzen-Connell effects.

The study of plant community dynamics has a long tradition. However, this field has barely incorporated the tools developed in the modern study of ecological networks. Key for this incorporation is the availability of a theoretical model able to incorporate field data about plant-plant interactions. In this study I introduce the Recruitment and Replacement (R&R) model that explicitly incorporates empirical networks of plant-plant interactions that occur during recruitment. The R&R model is built on fundamental demographic rates and incorporates competition for space between adults, intra- and inter-specific effects of established plants on recruitment and the colonization of vacant space. The basic analysis of the model provides predictions regarding different aspects of plant community dynamics, like the environmental conditions and species properties under which facilitation of recruitment is more likely to occur, the effect of recruitment facilitation on invasion, the effects of plant-plant interactions on equilibrium abundances and community stability, and the network properties that should relate to species equilibrium abundances. Many of these predictions agree with findings from published meta-analyses, supporting the general validity of the recruitment networks framework as a general approach to integrate the study of plant community dynamics into the study of ecological networks.