Pollinator declines and dependence on insect pollination, particularly in fruit and vegetable crops, creates a pressing need to understand growers' interactions with pollinators and factors affecting pollination strategies. At present, many growers are dependent on commercial honey bees (Apis mellifera), but diversified strategies may be necessary to secure adequate crop pollination in the future. As of yet, little social science research exists on pollination practices. This article presents the results of a survey of lowbush blueberry growers in Maine. The survey was part of a five-year pollination security study focusing on four fruit and vegetables crops in the Northeast US. The survey assesses grower perceptions of native pollinators' effectiveness and their perceptions of native pollinators' contribution to fruit set. Results indicate a widespread perception among growers of native pollinators' importance. While native pollinators are not effective enough to replace rented honey bees for three-quarters of the industry, they are broadly seen as an important form of insurance in poor weather when honey bees' effectiveness is reduced. The main obstacle to greater utilization of native pollinators found in this study was uncertainty over native pollinators' contribution to yield and the associated difficulty monitoring native pollinators' population size. Scientists and extension experts must work to reduce these obstacles before more widespread use of native pollinators will occur.

Recent pollinator declines highlight the importance of evaluating economic risk of agricultural systems heavily dependent on rented honey bees or native pollinators. Our study analyzed variability of native bees and honey bees, and the risks these pose to profitability of Maine's wild blueberry industry. We used cross-sectional data from organic, low-, medium-, and high-input wild blueberry producers in 1993, 1997-1998, 2005-2007, and from 2011 to 2015 (n = 162 fields). Data included native and honey bee densities (count/m2/min) and honey bee stocking densities (hives/ha). Blueberry fruit set, yield, and honey bee hive stocking density models were estimated. Fruit set is impacted about 1.6 times more by native bees than honey bees on a per bee basis. Fruit set significantly explained blueberry yield. Honey bee stocking density in fields predicted honey bee foraging densities. These three models were used in enterprise budgets for all four systems from on-farm surveys of 23 conventional and 12 organic producers (2012-2013). These budgets formed the basis of Monte Carlo simulations of production and profit. Stochastic dominance of net farm income (NFI) cumulative distribution functions revealed that if organic yields are high enough (2,345 kg/ha), organic systems are economically preferable to conventional systems. However, if organic yields are lower (724 kg/ha), it is riskier with higher variability of crop yield and NFI. Although medium-input systems are stochastically dominant with lower NFI variability compared with other conventional systems, the high-input system breaks even with the low-input system if honey bee hive rental prices triple in the future.