Net effects
We identified several three-way and four-way interactions, demonstrating how different effects combine to create net effects on species densities. For example, inquiline aphids can kill the nymphs of the galler B. dracunculifoliae (link 1), but because they preferentially occupy parasitised galls (link 21) and can kill parasitized nymphs (link 2), they can negatively affect parasitism and therefore also benefit the galler B. dracunculifoliae (link 15). As another example, aphid-tending ants, by increasing aphid inquilinism (link 11), possibly by leading aphids into galls, may positively affect parasitism on the galler B. dracunculifoliae (link 12) by reducing aphids on branches (links 14) and therefore their negative effect on parasitism (link 15). On the other hand, by increasing aphid inquilinism, ants negatively affect parasitoids, since parasitoid mortality is higher in galls occupied by aphids (link 2). Furthermore, hatched galls increase parasitism on the galler B. dracunculifoliae (link 16), perhaps serving as a clue for parasitoids, but they also increase aphid inquilinism (link 17) and therefore can reduce parasitism on the galler B. dracunculifoliae . Finally, parasitism of the galler B. dracunculifoliae was also positively associated with ant presence on branches (link 12), but the relationship was dependent on the presence of hatched galls (link 27; four-way interaction).
Every direct interaction on the network was influenced by at least one of the other groups. In several cases, it was not possible to propose a mechanism behind the indirect links, even though the natural history of the system is fairly well-known.