1. A fundamental question in community ecology is how the quantity of floral resources affects pollinator activity and how this relates to the structure and robustness of pollination networks. The issue has been mainly addressed at species level, while at community level several questions are still open.

2. Using a species-rich semi-natural grassland as model ecosystem, we explored the relationship between community-level quantity of floral resources, the number of flower visits at community-level and descriptors of the structure and robustness of pollination networks. To assess whether patterns of pollination interactions were consistent when using different types of floral resources, we measured three types of floral resources, namely the volume of nectar, the number of pollen grains and the number of flowers.

3. Community-level pollinator visitation rate showed a hump-shaped relationship with the quantity of floral resources. Results were consistent among models employing different measures of floral resources as independent variables. The relationship between the quantity of floral resources and the structure and robustness of the network was not consistent among different types of floral resources; only intermediate levels of nectar volume were associated with high levels of network robustness.

4. The pattern we found is in contrast with some former studies which reported positive linear relationships between the community-level visitation rate of pollinators and the quantity of floral resources. Rather, the observed hump-shaped pattern resembles the relationship previously found between the quantity of floral resources of a given plant species and its visitation rate per flower. Our results highlighted that, by using specific types of floral resources (e.g., the volume of nectar and the number of pollen grains) it is possible to reveal ecological processes that would be masked by using the number of flowers as the sole measure of floral resource.