Introduced species can have cascading effects on ecological communities, but indirect effects of species introductions are rarely the focus of ecological studies. For example, managed honey bees (Apis mellifera) have been widely introduced outside their native range and are increasingly dominant floral visitors. Multiple studies have documented how honey bees impact native bee communities through floral resource competition, but few have quantified how these competitive interactions indirectly affect pollination and plant reproduction. Such indirect effects are hard to detect because honey bees are themselves pollinators and may directly impact pollination through their own floral visits. The potentially huge but poorly understood impacts that non-native honey bees have on native plant populations combined with increased pressure from beekeepers to place hives in U.S. National Parks and Forests makes exploring the impacts of honey bee introductions on native plant pollination of pressing concern. In this study, we used experimental hive additions, field observations, as well as single-visit and multiple-visit pollination effectiveness trials across multiple years to untangle the direct and indirect impacts of increasing honey bee abundance on the pollination of an ecologically important wildflower, Camassia quamash. We found compelling evidence that honey bee introductions indirectly decrease pollination by reducing nectar and pollen availability and competitively excluding visits from more effective native bees. In contrast, the direct impact of honey bee visits on pollination was negligible, and, if anything, negative. Honey bees were ineffective pollinators and increasing visit quantity could not compensate for inferior visit quality. Indeed, although the effect was not statistically significant, increased honey bee visits had a marginally negative impact on seed production. Thus, honey bee introductions may erode longstanding plant-pollinator mutualisms, with negative consequences for plant reproduction. Our study calls for a more thorough understanding of the indirect effects of species introductions and more careful coordination of hive placements.

Field methods: In 2019, we selected 15 meadows, each >500 m apart, which varied in honey bee abundance across space and time due to their proximity to four seasonal apiaries. Across these meadows, we surveyed pollinator visitation patterns from May – July. To assess the overall abundance of honey bees and native bees, we netted active flower visitors while walking 100 m transects for two 30-minute periods (one between 8:00-12:00 and one between 12:00-16:00). To assess honey bee and native bee visitation rates to C. quamash, we conducted timed observations of 8-12 flowering C. quamash focal plants, noting all visitors in a 10-minute period. All focal plant observations occurred from 11:00-13:00, in-between morning and afternoon netting transects. At the end of each sampling day, we measured pollen and nectar availability in one to three flowers on each of 10-20 open-pollinated C. quamash plants and 10-20 unvisited control plants which were bagged on site arrival. We measured pollen availability as the proportion of dehisced anthers with pollen visible to the naked eye and measured nectar availability using 1µL capillary tubes.

Measuring pollination and seed set: At the end of each sampling day, we collected one style from a flower on twelve pre-marked C. quamash plants and mounted styles on fuchsin-tinted gelatin slides. We counted the number of conspecific and heterospecific pollen grains on stigmas using a compound light microscope. Seventy-two hours following pollinator observations and after the initiation of pollen tube growth, we collected a second style from these same pre-marked plants into 70% ethanol. In the lab, we softened styles with 8M NaOH at 35º C for 1 hour and stained pollen tubes by placing softened styles in a solution of 0.05% aniline blue in 0.1M KH2PO4 for 24 hours. We squashed styles beneath cover slips on microscope slides and counted pollen tubes near the base of styles using epifluorescence microscopy. Two weeks after conducting pollinator observations we collected fruits from these same plants and assessed seed set per fruit by scoring ovules as fertilized or unfertilized.

Single-visit effectiveness and controlled honey bee visit experiments: To assess the relative quality of honey bee visits and their direct contribution to C. quamash pollination, we performed two field experiments. In 2019, we conducted a controlled multiple-visit experiment to isolate the direct relationship between increasing honey bee visits and C. quamash pollination in one of the meadows where hives were introduced. In 2020, we returned to a different meadow, where honey bee abundance was more moderate, and assessed the single-visit pollination effectiveness of honey bees and other insect visitors. In both years, we bagged a selection of plants to prevent visitation and conducted observations from 7:00 until 16:00 across several days. For controlled honey bee visit experiments, we allowed a randomly assigned number of honey bee visits (between zero and twenty) to freshly opened flowers and all other visitors were excluded (i.e., swatted away). For single-visit effectiveness experiments, we allowed a single visit from different visitors, noting the pollinator identity and aspects of its visit behavior. Because we did not want to impact an insect’s visit by capturing it, our identifications were done in the field. We grouped visitors into several broad categories: Large-bodied Andrena spp., Apis mellifera, Bombus spp., Halictus spp., Osmia spp., “Small dark bees”, and syrphid flies. The category “Small dark bees” comprised primarily Lasioglossum spp. and small-bodied Andrena spp. For both field experiments, we re-bagged plants to prevent further visitation after experimental visits had concluded, collected fruits two weeks later, and counted fertilized ovules.

Data structure: please see metadata (included in the download as a README.md file) for more information about variable names and their interpretation.