SummaryThe basic units of ecological and evolutionary processes are individuals. Network studies aiming to infer mechanisms from complex systems, however, usually focus on interactions between species, not individuals. Accordingly, the structure and underlying mechanisms of individual-based interaction networks remain largely unknown.In a common garden, we recorded all interactions on flowers and leaves of 97 Sinapis arvensis individuals from seedling stage to fruit set and related interindividual differences in interactions to the plant individuals' phenotypes.The plant individuals significantly differed in their quantitative and qualitative interactions with arthropods on flowers and leaves. These differences remained stable over the entire season and thus were time-invariant. Variation in interacting arthropod communities could be explained by a pronounced intraspecific variability in flowering phenology, morphology and flower scent, and translated into variation in reproductive success. Interestingly, plant individuals with a similar composition of flower visitors were also visited by a similar assemblage of interaction partners at leaves.Our results show that the nonuniformity of plant species has pronounced effects in community ecology, potentially with implications for the persistence of communities and populations, and their ability to withstand environmental fluctuations.

Exotic species often cause severe alterations in native communities due to their ability to rapidly and efficiently utilize a broad spectrum of resources. In flower–visitor interactions, the breadth of resource use by native and exotic animals as well as the partitioning of resources among them is often estimated based on the number of (shared) plant species used as resources. However, whether a flower visitor is able to exploit plant resources has been shown to be delimited by functional floral traits such as morphological barriers or attractive or repellent chemical cues. Each of the ecologically relevant traits can be viewed as a dimension of a Hutchinsonian n‐dimensional hypervolume, which characterizes the range of phenotypes exploitable by a species. In this study, we quantified the sizes and overlaps of n‐dimensional hypervolumes defined by floral traits that are exploited by native and exotic flower visitors (afterwards referred to as exploited space, ES). In the heavily invaded Hawaii Volcanoes National Park, USA, we phenotyped 40 native and exotic plant species and recorded flower–visitor interactions. To quantify the size and overlap of ES, we applied dynamic range boxes. On average, exotic flower visitors were more generalized in resource use (larger ES) than natives ones, which is additionally indicated by the absence of native flower visitors on exotic flowering plant species. In particular, ES based on floral scent emission was larger for exotic flower visitors compared to native ones. The unevenly expanded ES of native and exotic animals led to an asymmetric overlap of floral ES where exotic flower visitors shared only a small proportion of their ES with natives but occupied a large proportion of the ES of natives. The asymmetry in resource use of native and exotic flower visitors suggests a potential advantage in resource exploitation of the latter, potentially explaining their success in Hawaiian ecosystems. Predicted range expansion of exotic plant and animal species may further increase the competition for and reduce the availability of resources for native animals. This may lead to further declines of native species and increasing threats for Hawaiian ecosystems. A http://onlinelibrary.wiley.com/doi/10.1111/1365-2435.12932/suppinfo is available for this article.