Deliberately increased network connectance in a plant-pollinator community experiment


Networks are a popular way to visually represent and analyse interactions between species in ecological communities, and there have been several recent calls for interaction network measures to be targets of conservation and management efforts. However, it is challenging to deliberately manipulate empirical network properties. Our objective was to determine whether the network property of connectance could be manipulated in a planned and deliberate fashion, without altering the size of the community. Connectance (proportion of realized interactions out of total number of possible interactions) is often used as a proxy for community complexity, and theory also suggests it is related to community robustness and stability. We designed a field experiment to increase the connectance of a mutualistic community by manipulating the attractiveness of plant species in a bee-plant interaction network. Specifically, we applied a chemical fertilizer to enhance the floral display and resource quality provided by experimental plants, and assessed the effect on bee community structure. We found the connectance was significantly higher in fertilized plots relative to controls. This manipulation was associated with a significant increase in both the species richness (+41%) and abundance (+77%) of pollinators; there were more pollinator species in treatment plots, and these pollinators visited more plant species with a higher frequency. This study shows that a network property can indeed be altered experimentally; here we were able to deliberately increase the connectance of a bee-plant network while keeping constant both the number of plant species and the background community of potential floral visitors. Deliberate manipulation of experimental community structure therefore could be a valuable avenue for future research. Integration of network theory with empirical research also has the potential to inform the design of network control approaches.

MIDAS Network Members