Field pairs were a minimum of 900 m apart to maintain independence . To avoid contamination of varieties, sunflower fields are moved every year; therefore no field was sampled in multiple years although two fields were adjacent to the same hedgerow in different years.In Yolo Co., acreage planted in sunflower has increased by over 55% during the past 5 years . It is the 8th most-planted crop in the region, grossing nearly $28 million USD in 2013 . It is produced mainly for hybrid seed, which is then grown for oilseed or confection. While sunflower is native to North America, the breeding system of sunflower grown for hybrid seed has been altered to be artificially gynodioecious, with separate male-fertile plants and male-sterile plants. For hybrid seed production, rows of male plants are interspersed with rows of female plants. Wild bees predominantly visit male plants to collect pollen for nest provisioning . Although honey bees visit both male and female plants, workers typically either collect nectar from female plants or pollen from male plants which limits cross pollination events . Honey bee movement between pollen and nectar producing rows of sunflower is often spurred by interference interactions with wild bees. When a wild bee and honey bee meet on a sunflower head, one or both fly to different sunflower heads or rows .
These interactions that increase pollen flow between rows also increase honey bee per visit efficiency, plastic seedling pots therefore have great potential to heighten seed set . Honey bees were stocked at an average rate of approximately 100 hives per field, or 1.5 hives per acre . We did not evaluate pest management because sunflower fields managed by different companies used similar practices. For example, all companies used pre-emergent herbicides prior to planting and seeds were treated with insecticides and either a fungicide or nematicide. Other management practices, including fertilization, tillage, row width and ratio of male to female rows, are also similar between companies , although irrigation practices vary by field, with the majority using furrow irrigation.Hedgerows were planted by growers to support beneficial insect populations, and include highly similar plant species composition . Hedgerows were 250–300 m long and 3–6 m wide. During the sunflower bloom period, only a portion of plants in the hedgerow were flowering . Eriogonum fasciculatum var. fasciculatum, Heteromeles arbutifolia, and Sambucus nigra ssp. cerullea were the only woody species in bloom. Forbs in bloom included Achillea millefolium, Asclepias californica, Asclepias fascicularis, Aster chilensis and Grindelia camporum. Weedy species were present in all hedgerows and most control sites; the dominant species were Convolvulus arvensis, Brassica sp., and Polygonum arenastrum. Control margins contained only non-native plant species, or were maintained as bare, weed-free areas. Bare/weedy field margins in our study region are managed by burning, herbicides, or scraping; no management actions took place during our study period.
By design,hedgerows contained more plant species and more blooms than control weedy edges .To quantify the landscape surrounding each site we created 18 land use categorizations . We then hand digitized National Agriculture Imagery Program within a 1 km buffer around study sites in ArcGIS 10.1 . To determine landscape effects on wild bee populations in sunflower, we examined the proportion of habitat within each buffer that could provide resources to wild bees . This included both natural habitats and altered habitats . Potential pollinator habitat around our study sites varied from 1 to 40%, with a median of 5% . Control and hedgerow sites were paired by landscape context to minimize differences.We established two 200 m transects within each field, perpendicular to the field edge or hedgerow and 100 m apart . We netted and observed pollinators at four distances along these transects: 10, 50, 100 and 200 m from the edge. We varied the starting sampling location within fields and edges at each study site to reduce conflation of distance with temporal variation in bee foraging behavior, which peaks in the morning and late afternoon . Each site was sampled once, during peak bloom , on a clear day with wind speeds <2.5 m/s and temperatures >18 C between 08:00 h and 14:00 h. We visually observed visitation for 2 min each in two male fertile and two male-sterile 2 1 m plots at each distance. Within hedgerows and edges we haphazardly sampled floral visitors for 2 min in eighth plots containing floral blooms. Only insects that contacted the anthers or stigmas were recorded as floral visitors. We also recorded non-bee visits; these accounted for <1% of all visits and were, for simplicity, excluded from analyses. We were unable to identify bees to species in visual observations; therefore we classified them to citizen science categories from Kremen et al. of all records. We did not include feral Apis in our wild bee categorization because we were unable to distinguish them from managed Apis.To determine ambient pollination rates, we collected three sunflower heads at each distance/transect combination prior to harvest. In the first year of this study we bagged one male-sterile sunflower head at each distance along both transects to determine seed set levels without cross-pollination events.
No seeds were produced in any bagged sunflower head, therefore we did not account for seeds set due to selfing in our models of seed set. Heads were dried, measured, and all mature seeds were removed, weighed and counted with a Syntron automatic seed counter. We tested for differences in head size between companies using a generalized linear model, with site nested within pair as a random effect, in the R package lme4 . Sunflower head size was similar between companies , although one company had a wider range of head sizes and was significantly different from the other three companies in the study . All hedgerow and control sites were paired by company.Sunflower specialists are more effective pollinators of sunflower than generalist species . We therefore also investigated whether sunflower specialists were more abundant in hedgerow or control field edges using an independent data set from 26 hedgerows and 21 control edges in Yolo Co. . Floral visitors were netted for 1 h in hedgerows and control edges during 4–5 sample rounds between April and August in 2012–2013. We queried this specimen database for sunflower specialist bees, which we defined as primary oligoleges . To assess whether the amount of nearby sunflower in the landscape impacted sunflower specialist presence in field edges in the independent dataset, we constructed 1 km buffers around sites in ArcGIS 10.4 and recorded the proportion of sunflower fields around each site using pesticide spray records , which identify which crop is grown on each parcel, and the California crop improvement sunflower isolation map .We used a chao estimator to evaluate species richness within sites in the R package vegan . To determine the impact of hedgerow presence, field location , and surrounding pollinator habitat in the landscape on wild bee species richness and abundance we used general linear models with Poisson and negative binomial distributions respectively in the R package lme . Both models included an interaction between hedgerow presence and field location. We used raw species richness because we only sampled each site once and some sites contained too few individuals for estimation or rarefaction . We also assessed factors influencing sunflower visitation rates by honey bees and wild bees. Hedgerow presence, distance from hedgerow, and their interaction, potential pollinator habitat and sunflower sex were independent variables. In species richness, abundance and visitation models, site nested within pair was included as a random effect. We evaluated the differences between the community of bees in control edges, hedgerows and crop fields using a per MANOVA on their Chao1 dissimilarities in the R package vegan . We then determined whether male and female sunflower specialist bees utilized hedgerows or control field edges using the independent data set . We modeled counts of bees as the dependent variable with a Poisson distribution in the R package lme4 . Hedgerow presence, proportion of sunflower and potential pollinator habitat within a 1 km radius, bee specialization on sunflower, bee sex and an interaction between specialization and hedgerow presence were the independent variables. Site nested within pair was included as a random effect. To determine which factors impacted sunflower seed set, containers size for raspberries we used negative binomial generalized linear models in the R package lme4 that accounted for overdispersion in the seed data .
We examined the effect of wild bee abundance and richness on seed set from net and visitation data separately. We used raw species richness because some site distance combinations contained too few individuals for estimation or rarefaction . In all models, sunflower seed set was the dependent variable. In the model for netted bees, independent variables were hedgerow presence, wild bee abundance, wild bee species richness, sunflower company, distance into the field from the edge, and an interaction between netted wild bee abundance and honey bee visitation . For the model including visitation rates, additional explanatory variables included aggregate wild bee visitation to male-fertile and male-sterile flowers, honey bee visitation, and an interaction term between wild bee visitation and honey bee visitation. All continuous variables were scaled /sd). We checked all variables for collinearity , and no collinear variables were included in any model. For example, sunflower head size was correlated with variety. However, varieties were specific to sunflower company, so only sunflower company was retained in the model.Measuring the levels of ecosystem services derived from field edge habitat management in a variety of contexts is critical to demonstrating their efficacy and flexibility. If services are highly variable over time or from site to site, costs may outweigh the benefits and limit the adoption of diversification practices . Although other studies have found that field-edge diversification increase pollinator populations both in edges and fields and enhance pollination services to crops in adjacent fields , we did not detect any differences in rates of seed set in sunflower fields adjacent to hedgerow or control edges. Wild bee richness and an interaction between wild bee visitation and managed honey bee visitation, however, positively impacted seed set; yet these factors were not influenced by hedgerow presence. Proportion of pollinator habitat in the surrounding landscape did not influence the bee community visiting sunflower, despite a large body of evidence supporting strong positive landscape effects . We did find higher numbers of sunflower specialist bees in hedgerows than in control sites. Based on these findings, we conclude that sunflower in not a good candidate crop for field edge enhancements, at least in our study region, although they exhibit potential for supporting populations of sunflower pollinating bees. We detected distinct differences in community composition of wild bees present in edges versus fields. This difference was likely driven by the fact that the dominant bee species found within fields, sunflower specialists, were either rare visitors to or absent from both hedgerow and control edge habitats. We only sampled each site once, therefore increased sampling could lead to more convergence or divergence between bee communities in these habitats. There can be significant overlap between species found in MFC fields and adjacent hedgerows , however species composition in hedgerows has also been shown to more closely resemble bee communities in forest habitat than adjacent crop fields . One factor likely driving the differences in species composition in our study region is the absence of sunflower planted within hedgerows due to concerns about genetic contamination of sunflower crop varieties. Because female sunflower specialists collect only sunflower pollen to provision their nests, they may not be attracted to the resources in hedgerows during the sunflower bloom period, instead being drawn into fields . Nevertheless, assessment of the independent dataset indicated that hedgerows provide important floral resources to sunflower specialist bees, especially males. Male sunflower specialists have been observed investigating honey bees as potential mates, which increases honey bee movement from male-fertile to male-sterile sunflowers and increases their pollination efficacy . Male bees, therefore, likely contribute to the interactive effect between wild bee richness and honey bees on rates of seed set. We found a slight positive effect of wild bee species richness on seed set rates, indicating that a higher number of bee species benefits pollination function in sunflower. Functional complementarity between species can enhance fruit and seed production in a variety of crops . Bee foraging behavior and bee body size can influence within inflorescence foraging, leading to more complete pollination in a single flower .