Background/Question/Methods Recent theoretical models have explored the implications of non-trophic interactions for community assembly processes by allowing both direct consumptive links between species and interaction modifications (IMs, whereby the direct consumptive link between two species is modified by a third party). These models suggest that IMs may be important largely because they alter mean consumption rates and connectivity among members of the regional species pool from which communities are assembled. However, the models also employ assumptions that constrict the potential effects of IMs on consumptive effect strengths and connectivities, leaving the generality of the results unclear. Here, we compare the results of community assembly models with IMs defined by a variety of functions and with various patterns of potential direct consumptive links between members of the regional species pool. We use these models to explore the degree to which IMs either have consistent effects, or effects that are consistently understandable in terms of the changes in consumption rates and connectivities that the IMs cause, across models. Results/Conclusions IMs generally reduced species richness and diversity of the steady-state communities present at the end of the community assembly process. However, the effect depended both on the assumptions governing the IMs and on the pattern of direct linkages in the species pool. Surprisingly, functions for assigning IMs leading to lower mean consumptive strengths, which would have been expected to have been stabilizing, often had more negative or less positive effects on species richness and diversity than functions leading to higher mean strengths. The underlying pattern of consumptive links were also important: IMs had stronger and more consistently negative effects when consumers fed both on species one trophic level down and on species within their own trophic level than when species fed on prey one and two trophic levels down or when species only fed on prey one trophic level down (which was the only case where positive effects of IMs were seen). Within this last scenario, the effects of IMs also often became stronger as underlying trophic connectance increased. These models show that IMs' effects in large communities may be more nuanced than previously appreciated, but also suggest that there are some generalities concerning these to be found.
