Beneficial insects provide valuable services upon which we rely, including pollination. Pollinator conservation is a global priority, and a significant concern in Ireland, where over half of extant bee species have declined significantly in recent decades. As flower-visiting insects rely on flowering plants, one way to conserve and promote pollinator populations is to protect high-quality habitat. We analysed the structure of insect-flower interactions from multiple habitat categories in a large database of interactions from Ireland. Our primary goals were to compare spatial and temporal variation in Irish network structures, compare Irish networks to published networks from other countries, and provide evidence-based recommendations for pollinator conservation in Ireland by identifying well-visited plant species that may promote high pollinator diversity, abundance, and functional complementarity. Habitat types within Ireland differed substantially: semi-natural grasslands had the highest pollinator species richness and largest number of unique pollinator species, while intensively-managed habitats exhibited negative asymmetry (more plant than pollinator species). This negative asymmetry is notable because most plant-pollinator networks exhibit a positive asymmetry. Within intensively-managed habitats, agricultural and urban habitats differed. Urban habitats had the highest number of non-native plant species while agricultural habitats had the lowest pollinator species richness. We also found Irish networks varied across the growing season, where July had the highest plant and insect species richness. When comparing Irish networks to published networks from other countries, we found Irish networks had a higher ratio of plant species to pollinator species, and that this difference was most evident in agricultural habitats. This ratio means the typical network asymmetry (more pollinator than plant species) was flipped (more plant than pollinator species) in the Irish network. We conclude that conserving semi-natural grasslands in Ireland will be an essential component of pollinator conservation and identify thirty-five plant species important for restoring semi-natural habitats.

The data used for these analyses were collated from six different studies in Ireland where obligately flower-visiting insects (specifically Lepidoptera, bees, and hoverflies) were observed/collected while foraging on flowers. The surveyor recorded both insect and plant identification to the highest resolution possible and took samples where field identification was not possible. These interactions were recorded from May through August in 2009-2011 and 2017 and April through October in 2018. The surveyor assigned the habitat type for each sampling event according to the Fossitt guide to habitat types of Ireland. In most cases, the surveyor also quantified the floral abundance of the different plant species visited during a transect. Altogether, these constituted 800 surveys and ~ 940 hours of sampling time over more than 709,721 m² of transect area. We observed/collected 4,538 specimens representing 148 insect visitor species on 239 plant species. Of the plant species identified in this study, 74 were not native to Ireland.

From these datasets, we automatically excluded any data where the host plant or insect visitor was not resolved to at least genus. We included some specimens identified to genus only, as some plant and insect taxa were difficult/impossible to resolve to species, often listed as “aggregates” (e.g. Taraxacum agg. and Bombus lucorum agg.). Only one study reported plant species present that did not receive visits (“true zeroes”), so we excluded these records. For the phenological comparisons, we included all remaining data, but when comparing habitat types, we excluded records where the surveyor included more than one habitat type in a single survey. When ranking species importance, we only included data where the surveyor counted the number of floral units (e.g. inflorescences, racemes, umbels) to quantify visitation rate (number of flower visitors per unit floral area).

Changes after Sep 21, 2022: 

Corrected errors in the latitude/longitude