The metacommunity perspective appears as an appropriate conceptual framework to make ecology more predictive. It is particularly relevant to limnology, where exchanges of organisms and nutrients affect community and ecosystem properties from the local to the regional scales. The recent development of neutral theory appears as a step back in that direction because of the assumption of ecological equivalence and the absence of any effect of the environment on community organization. A remarkable strength of neutral theory is nonetheless to provide a general theory of diversity that accounts for a wide range of empirical observations. In this paper, we argue that neutral theory can be useful to understand the impact of dispersal on community assembly in landscapes of various complexities. Our analysis focus on spatially explicit landscapes conceptualized as networks of local communities (e.g., lakes) connected to each other by dispersal channels (e.g., rivers). The main objective of the paper is to use neutral theory to stress the importance of landscape structure on the distribution of diversity. We refer to the landscape organization as a spatial contingency that could potentially affect the coexistence mechanisms at play. We briefly review the main approaches to describe spatial networks and describe three simple toy models of metacommunity dynamics. We take this opportunity to review their assumptions and main predictions. We then conduct simulations of these models to reveal with simple examples the impact of spatial network structure on diversity distribution. The simulation results show that competitive interactions buffer the potential impact of landscape structure. The strongest relationship between node position in the landscape and species richness was observed for the patch dynamics model without any interactions. On the other hand, strong and unequal competitive interactions minimized the effect of node position. We conclude that the neutral model is a useful tool to understand the joint effects of dispersal and ecological interactions. Our analysis shows that limnologists must now integrate more realistic landscapes when analyzing community assembly from a metacommunity perspective.