Floral visitation alone has been typically used to characterize plant-pollinator interaction networks even though it ignores differences in the quality of floral visits (e.g. transport of pollen) and thus may overestimate the number and functional importance of pollinating interactions. However, how network structural properties differ between floral visitation and pollen transport networks is not well understood. Furthermore, the strength and frequency of plant-pollinator interactions may vary across fine temporal scales (within a single season) further limiting our predictive understanding of the drivers and consequences of plant-pollinator network structure. Thus, evaluating the structure of pollen transport networks and how they change within a flowering season may help increase our predictive understanding of the ecological consequences of plant-pollinator network structure. Here we compare plant-pollinator network structure using floral visitation and pollen transport data and evaluate within-season variation in pollen transport network structure in a diverse plant-pollinator community. Our results show that pollen transport networks provide a more accurate representation of the diversity of plant-pollinator interactions in a community but that floral visitation and pollen transport networks do not differ in overall network structure. Pollen transport network structure was relatively stable throughout the flowering season despite changes in plan and pollinator species composition. Overall, our study highlights the need to improve our understanding of the drivers of plant-pollinator network structure in order to more fully understand the process that govern the assembly of these interactions in nature.

To sample the pollinator community four 1x40m transects were set up at the study site. Pollinators were sampled by walking each transect at a constant pace collecting all insects observed visiting flowers until a maximum of 60. Sampling took place between May and August 2019 during peak pollinator activity between 8:00AM and 6:00PM for 21 days across 13 weeks. All insects were collected with butterfly nets when they were observed making contact with the floral reproductive structures (anthers and stigma). Insect pollen loads were sampled by swabbing the body of each floral visitor collected with fuchsin jelly cubes. The corbiculae of bee species was avoided as the pollen located within it is not typically available for pollination. To account for ‘incidental’ pollen transport we only considered PT interactions where an average of five or more pollen grains of a specific plant species where found on a floral visitor. We further improved the reliability of our estimate by only removing interactions of ≤ 5 pollen grains that also constituted less than 5% of the total pollen load carried by each insect species. We evaluated PT network structure across early (ES), middle (MS), and late flowering season (LS) by constructing pollen transport networks for each individual week (13 total weeks) and then categorizing each week as early, middle or late flowering.