Non Technical Summary
Many factors affect the stability of wild bee populations, including land-use change and competition with non-native species, such as honeybees. Wild bees are often equal or superior to managed honeybees in their ability to provide pollination services, yet many hobbyists, producers, and special interest groups advocate for 'backyard beekeeping' and supplementation of pollination in both natural and agricultural systems with honeybees. This may have consequences for the sustainability of ecosystem services if honeybees compete, either directly or indirectly, with wild bee populations for access to limiting pollen resources. Currently, there exist no clear policy recommendations concerning honeybee hive density in landscapes of the American west. In addition, supplementation with honeybees may be inefficient for producers if fees are paid for hive deployment when wild bee populations provide adequate pollination services. Accordingly, there is a need to understand (1) the extent and strength of resource overlap and competition between wild bee populations and managed honeybees, (2) determine whether honeybees affect nest-site use and reproduction of wild bees, and (3) quantify the relative efficacy of wild bees in providing pollination to crops. We propose here a set of experiments to characterize interactions between wild bees and honeybees, interpreting interactions in the context of landscape variation, with the goal of providing new information on which wild bee taxa are likely to be impacted by honeybees, and why. Our research will apply foundational principles in insect ecology to determine whether honeybees have deleterious effects on wild bee foraging and vital rates in a peri-urban landscape with a rapidly growing wildland-urban interface. Resulting datasets and extension products will inform future policy recommendations on honeybee stocking densities in the wildland-urban interface, and can also help growers to make decisions about factors affecting the pollination of several regionally important field crops in the Rocky Mountain region.

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
136	3085	1130	75%
211	3085	1130	25%

Knowledge Area
136 - Conservation of Biological Diversity; 211 - Insects, Mites, and Other Arthropods Affecting Plants;

Subject Of Investigation
3085 - Other bees (non-Apis bees) - All bees except Apis cerana; includes Bombas, Alfalfa leaf-cutter;

Field Of Science
1130 - Entomology and acarology;

Goals / Objectives
Our overarching goal is to evaluate competition between managed honeybees and wild bees. Focusing on both direct (bee-bee interactions) and indirect (resource-mediated) interactions, as well as patterns of habitat use and pollination effectiveness, the objectives of the proposed work are to:Characterize competition (interference and exploitation) between managed honeybees and wild bees relative to honeybee hive density and landscape context,Evaluate nest-site use and reproductive success of wild bees in the presence and absence of honeybees, andAssess effectiveness of wild bees as pollinators of regionally important crops.

Project Methods
Objective 1) Characterize competition (interference and exploitation) between managed honeybees and wild bees relative to honeybee hive density and landscape contextWe will experimentally deploy honeybeecolonies at varying densities and across multiple landscapes (n=36 sites) to evaluate changes in pollination network structure before, during, and after hive introduction to landscapes. Pollination networks will be analyzed by netting bees from flowers, and identifying all bees and flowers in this process to characterize which bees interact with which flowers.At each site, bee visitations (by wild bees and A. mellifera) to flowers will be sampled within a 2-ha fixed-area plot. Records of visitations will subsequently be used to construct bipartite interaction networks to analyze effects of honeybees on bee-flower networks. Each sampling bout will last for 1 hour, where two observers will randomly walk the plot to simultaneously record bee visitations to flowers.The foraging time (seconds) of bees observed alighting on flowers will be timed with a stopwatch. Apis mellifera is identifiable on sight and will not require capture for identification, but wild bees that cannot be immediately identified to genus will be netted and visually identified to the lowest taxonomic level before being released (genus and in many cases species). A subset of foraging bees (every 5th bee [20% of the sample], including those with visibly evident pollen loads on scopa or corbicula) will be collected into vials and kept for analysis of pollen mass and wing wear. Wing wear will be analyzed by examination of specimens under the microscope.For each interaction, observers will record (1) bee taxa (genus or species), (2) foraging time (seconds), (3) distance to hive box (if present; using a handheld rangefinder), and (4) a binary observation of whether honeybees alighted on flowers during the foraging observation period. The study will be repeated over two growing seasons and this sampling design will yield a dataset of foraging times, foraging efficiency (pollen collected and wing wear), and the total resource set foraged upon by regional bee taxa as well as A. mellifera. A variety of statistical models (GLMMs) will be used to rank factors predicting both interference and exploitation competition, and their effects on bipartite network structure variables of interest including modularity, specialization, and nestedness. We will also model the degree of nich overlap (in terms of foraging resources) between native bees and honeybees.Objective 2) Evaluate nest-site use and reproduction by wild bees in the presence and absence of honeybeesWe will source O. lignaria from commercially available populations in the first year of the study; in subsequent years we will use cocoons from overwintered individuals. At each site, we will place two 32-hole wooden nesting blocks, with holes lined with paper straws (Jonesville Paper Tube Corp., Jonesville, MI). Pre-sorted male and female cocoons are placed in a 3:2 ratio into 1" × 5" PVC 'release tubes' (n=40 bees released from each tube; 24 males and 16 females) that are capped but have a small hole drilled in one end and affixed to the base of each nesting block (2 per site). Tubes are placed in early May, and as temperatures warm bees emerge naturally from tubes, with 30-50% of released individuals returning to utilize nesting blocks. In ongoing studies in the region, we have found that other cavity nesting bee taxa, especially other Osmia species, also make use of nesting blocks. Nesting straws will be monitored weekly by marking nest development on the straw and removing pollen from small incisions in the paper straw. At the end of each season, all collected nests will be stored in darkness at 22°C for 6 months, followed by 4 months at 6°C to overwinter. We will use X-ray photography to determine the number, sex, and size of all offspring from nest blocks. We will also collect and rear any wild cavity nesting species that establish within blocks.Because commercially available Bombus species are unavailable for outdoor use in our region, we will rear B. griseocollis from wild-caught queens following protocols described previously by co-PI Mola. Bombus griseocollis is a locally abundant species proposed for commercialization and favorable to captive rearing with an establishment success rate of 76% in a recent study. After rearing, two colonies will be placed at each site early in the season. Colonies will be weighed every 10 days, and pollen collected from returning foragers using non-lethal techniques. After decline of colony weight (a sign of a shift to the production of reproductives instead of worker castes), colonies will be removed from the field and preserved via freezing to allow for counts of total cocoon and queen production.We will also analyze pollen collected from bee nests. Pollen will be suspended in water on a microscope slide and then heated to allow grains to expand and evaporate excess water. Fuchsin jelly will be used to stain and aid in identification. Pollen species will be accounted along three randomly selected lines across the cover slip at 400x magnification. Species representing less than 2% of the load will be excluded as they likely represent contamination. Pollen grains will be identified using reference material from the literature as well as collections of single-species reference slides conducted during field work. If necessary, acetolysis will be used to aid in identification. \We will use a linear mixed modeling framework to test a variety of hypotheses related to factors predicting nest site use, bee fitness traits, and pollen collection efficiency (and foraging breadth).Objective 3) Assess effectiveness of wild bees as pollinators of regionally important cropsStudies of pollination effectiveness will use greenhouse-grown plants (alfalfa, sunflower, and cantaloupe) deployed in field locations with and without honeybee colonies; honeybee treatments (0, 1, or 3 colonies at a site, as in the above tests) using a 'pollination deficit' treatment.Plants will be grown in the greenhouse in 4-L pots with a standard potting soil mixture (Promix) containing mycorrhizae and amended with a slow-release fertilizer (Osmocote) and perlite. The timing of sowing in pots will be matched with typical field planting dates so that test plant bloom phenology approximately matches that of plants grown under field conditions. In the proposed study region, the first flowering of field alfalfa typically occurs in late May through early June but may then be relatively continuous at some sites, whereas sunflower and cantaloupe flower in mid-July and late July/early August, respectively. Thus, tests of pollination efficacy will be staggered for the three selected study species, allowing for researchers to handle and attend to only one crop species at a time. Tests will quantify both (1) average yields (total seed weight per plant for alfalfa and sunflower and total fruit weight per plant for cantaloupe) as well as (2) the 'pollination traits' of floral visitors during each trial. Trait assignments and their relative frequencies in the assemblage of floral visitors will subsequently be used to explain variation in yield, in addition to landscape variation in honeybee density.