Pollinator-dependent crops rely on the activity of managed and wild pollinators. While farm management and surrounding landscape can influence wild pollinator contributions, managed pollinator contributions may be primarily driven by their stocking densities, though this is not well studied across crops. We selected 20 southern highbush blueberry farms along two independent gradients of honey bee Apis mellifera L. (Hymenoptera: Apidae) stocking density (~1–11 hives/acre) and bumble bee Bombus impatiens Cresson (Hymenoptera: Apidae) stocking density (0 - 3 colonies/acre) ensuring that stocking densities were not correlated with farm or landscape attributes. Across farms, we observed managed and wild bee visitation rates, and measured yield estimates. Farms with greater bumble bee stocking densities had higher bumble bee visitation rates and yield estimates, but farms with higher honey bee stocking densities only received higher honey bee visitation rates at the end of bloom and did not have higher yield estimates. The main wild pollinator, the southeastern blueberry bee Habropoda laboriosa (Fabricius) (Hymenoptera: Apidae), showed higher visitation rates on organic farms and in late bloom. In general, higher visitation rates by honey bees, bumble bees, and H. laboriosa were correlated with higher yields. Our results suggest that yields are limited by bee visitation rates, and that within the stocking density ranges studied, increasing managed bumble bees, but not honey bees, increases their visitation rates. While H. laboriosa had the greatest effect on yield estimates, its activity appears to be limited by both a phenological mismatch with crop bloom and farm management.

In cucurbit crops such as watermelon, implementation of integrated pest management (IPM) is important due to the high reliance on bees for fruit set, along with mounting evidence of the risks of insecticide use associated with pollinator health. Yet, IPM adoption, on-farm pesticide use behaviors, their costs, and impacts on the primary insect pest (striped cucumber beetle, Acalymma vittatum F.) are poorly known in one of the key watermelon-growing regions, the Midwestern United States. To better understand how to implement IPM into watermelon production, we assessed pest management practices on commercial watermelon farms using 30 field sites in Indiana and Illinois over 2 yr in 2017 and 2018. Across all sampling dates, beetles never crossed the economic threshold of five beetles/plant at any farm and most were maintained at densities far below this level (i.e., <1 beetle/plant). Moreover, we documented a wide range of insecticide inputs (mean ca. 5 applications per field per season; max. 10 applications) that were largely dominated by inexpensive foliar pyrethroid sprays; however, insecticide application frequency was poorly correlated with pest counts, suggesting that most of these applications were unnecessary. We calculated that the cost of the average insecticide program far exceeds the cost of scouting, and thus IPM is estimated to save growers ca. $1,000 per field under average conditions (i.e., field size, insecticide cost). These data strongly indicate that current management practices on commercial farms in the Midwest would benefit from implementing more threshold-based IPM programs with potential increases in both farm profitability and pollination services.

Pesticides threaten ecosystem services by reducing the abundance and diversity of beneficial arthropods, including pollinators, in agroecosystems (Carvalho, 2017). Pesticide use can result in hazards to honeybees Apis mellifera L. and wild bee species, and is considered a factor contributing to pollinator decline (Zioga et al., 2020). These non-target effects reduce crop visitation, disrupt pollination and can reduce yields (Stanley et al., 2015). However, the impacts of pesticides on pollinators are rarely studied beyond the focal field, or local level, despite the fact that some bees forage widely (Greenleaf et al., 2007) and thus pesticide exposure occurs at a larger spatial

Insecticide use and insufficient forage are two of the leading stressors to pollinators in agroecosystems. While these factors have been well studied individually, the experimental designs do not reflect real-world conditions where insecticide exposure and lack of forage occur simultaneously and could interactively suppress pollinator health. Using outdoor enclosures, we tested the effects of insecticides (imidacloprid + lambda-cyhalothrin) and non-crop forage (clover) in a factorial design, measuring the survival, behaviour and performance of bumblebees ( Bombus impatiens ), as well as pollination of the focal crop, watermelon. Colony survival was synergistically reduced to 17% in watermelon alone + insecticides (survival was 100% in all other treatments). However, behavioural shifts in foraging were mainly owing to insecticides (e.g. 95% reduced visitation rate to watermelon flowers), while impacts on hive performance were primarily driven by clover presence (e.g. 374% increase in the number of live eggs). Insecticide-mediated reductions in foraging decreased crop pollination (fruit set) by 32%. Altogether, these data indicate that both insecticides and non-crop forage play integral roles in shaping pollinator health in agricultural landscapes, but the relative importance and interaction of these two factors depend on which aspect of ‘health’ is being considered.