Accounting for the identity of species shows that pollinator communities in different farm fields across large areas basically consist of variations of the same core set of species that prefer to forage on crops and that are augmented with the occasional new species. So while there is little doubt that a reduction in the local diversity of crop-visiting bee species may have negative consequences for the pollination services they deliver, here we show that even the cumulative number of species across species-poor and species-rich fields represents only a small proportion of all bees and are dominated by an even smaller subset of species that occur on most fields . One benefit of biodiversity to ecosystem services is that it may provide insurance effects that stabilize services over time or space. Our results are in line with this because for most bee species large contributions to pollination were limited to specific years, crops and/or sites . It could therefore be argued that in order to maintain stable pollination services, one would need to conserve a much wider set of bee species than those that are currently numerous on crops. Species that are now rarely observed may, after all, become important in the future. While this may be true, this line of reasoning only applies to bee species that can actually use crop plants for forage. Bee species, even generalists,cut flower bucket have distinct preferences for host plants and may be incapable of raising offspring on resources from non-preferred plants such as agricultural crops .

Species preferring non-crop plant families show more negative population trends than species specializing on members of crop plant families, thereby confirming that many bee species fail to make use of this abundant resource supply. Thus, many of the bee species that are currently absent from crop flowers are unlikely to be important for spatial or temporal insurance effects of pollinator biodiversity on crop pollination, simply because they will not utilize crops even if conditions change. Many previous studies have found that species richness of bee communities in agricultural landscapes declines with decreasing proportion of semi-natural habitats. Our findings present a novel and more nuanced interpretation: while most bee species decline in abundance with expansion of agriculture, the species currently providing most of the pollination services to crops persist . Previous studies on plants have likewise demonstrated that with increasing land use intensity subdominant species are the first to decline, whereas dominant species are little affected. Whether bee communities consisting of only the dominant pollinators are capable of providing sufficient pollination is unclear, but this pattern suggests that land use change will affect crop pollination less than it affects biodiversity. Measures to mitigate loss of pollination services are most cost effective in relatively intensively farmed landscapes because here measures have the highest impact, ecosystem service delivery is likely to be reduced owing to the intensive farming practices, and returns on investments are greater owing to higher yields in intensively farmed areas.

Our results show that pollinator habitat creation in intensively farmed landscapes can successfully enhance the dominant crop-visiting bee species , but are unlikely to benefit threatened species because of lack of source populations. Species are classified as threatened when their numbers have experienced significant declines or their geographical distributions have contracted. Agricultural intensification is an important driver of species decline1. It is therefore perhaps not surprising that, in agricultural landscapes, threatened species contribute little to ecosystem service delivery, and benefit little from general conservation measures. However, in the past, many of the species that are now threatened occurred widespread and contributed to pollination services on more extensively managed farmland. Threatened species may also still dominate bee communities in restricted parts of their former distributional range. Effective conservation measures for threatened species should therefore be targeted towards these bee species and their habitats, and not the crops to be pollinated. Highlighting the economic benefits people might obtain from biodiversity can be an effective instrument to motivate people or institutions to support biodiversity conservation. However, too much focus on the services delivered by pollinators may lead to adoption of practices that will not benefit species that could potentially contribute under changing agricultural conditions nor species that will never contribute to crop pollination. Benefits of biodiversity should therefore not be used as the sole rationale for biodiversity conservation as, for example, is currently done in the new strategy of the Convention on Biological Diversity7 and in the EU biodiversity strategy to 2020 .

Moral arguments remain pivotal to supporting conservation of the larger portion of biodiversity including threatened species that currently contribute little to ecosystem service delivery. Such arguments are powerful and define many human actions, from taking care of the elderly to preserving historical buildings or art. Ecologists and conservationists need to make these distinctions clear if we expect policy makers or land owners to defend species with no clearly defined economic value to humans.Diapause is a strategy that many species of insects rely upon to avoid adverse environmental conditions. For example, in temperate regions, insects typically enter diapause to better cope with freezing temperatures . Although bypassing environmental stressors via dormancy has obvious fitness benefits, diapause can still be an energetically and physiologically taxing life cycle stage . Because they are unable to replenish energy stores, insects face potential damage from cold shock, desiccation, and depletion of fat reserves during diapause. These kinds of damage can result in reduced survival and post-diapause fitness . There is concern that some insect species may be particularly vulnerable to mortality during diapause, and that low overwintering survival rates may lead to “demographic bottlenecks” , where conditions experienced during the winter disproportionately impact population dynamics. Though many species of insects spend most of their lifespan in diapause , it is often unclear how sensitive natural populations are to overwintering mortality. To help address this knowledge gap, we conducted a field-based study to estimate diapause survival of queen bumblebees . Many studies that address insect diapause are carried out entirely in the lab, often with the goal of characterizing the physiological mechanisms that underpin survival . Researchers may rely on lab based methodologies, at least in part, because studying the ecology of overwintering insects in the field can be incredibly challenging. Many insect species spend the winter buried in the soil or under leaf litter or woody debris,flower display buckets and are thus difficult to locate and monitor. Monitoring vital rates in the lab also allows researchers to better control sources of environmental variation. However, the conditions experienced by animals in artificial settings are not always ecologically realistic. For this reason, it is often unclear if discoveries about dormancy from lab-based studies can be extrapolated to natural populations . Perhaps unsurprisingly, lab-based methodologies also play an essential role in studies of the biology of hibernating bumblebees. Bumblebees are some of the most important native pollinators of crops in the United States , and recently there has been great interest in understanding their biology throughout the life cycle . Multiple studies have monitored diapausing bumblebees in the lab, relating survival rates to environmental factors like temperature or relative humidity or to intrinsic physiological factors like body mass . Many of the studies that target this life cycle stage are focused on improving methods of husbandry or commercial rearing . Conditions experienced by queens in the lab are not always ecologically realistic . For example, researchers may monitor overwintering queens for only 1–4 months , even though bumblebee queens in temperate regions generally spend 6–9 months in diapause . From lab-based studies, we know that maintaining queen bumblebees during diapause can pose a significant hurdle in rearing captive populations in the lab . However, we do not know whether low rates of diapause survival observed in lab studies are typical in nature .

In this study, we estimate diapause survival rates of queen bumblebees in the field by monitoring queens overwintering in large aggregations. To compare our estimates of diapause survival to those produced by lab-based studies, we conducted a meta-analysis of past research, using Web of Science and OATD to build a data set using 32 different studies of bumblebee survival in laboratory conditions. In addition to providing general information regarding the locations of overwintering B. impatiens queens, our study is the first to estimate overwintering mortality of natural bumblebee populations .Bombus impatiens is one of the most common species of bumblebee in the Northeastern United States , including in our study region . The primary reason we chose to work with B. impatiens is because the overwintering ecology of this species is well suited to field-based monitoring. Most bumblebees overwinter below the ground , and like most other social Hymenoptera, overwinter solitarily . However, there are a handful of social bee species that overwinter in large aggregations in the soil around natal nest sites . Thus, although locating queen bumblebees in sufficient numbers has posed a major hurdle to studying diapause for other bumblebee species in the field , it is feasible to monitor large numbers of overwintering queens in B. impatiens as long as nest sites can be found. B. impatiens is also one of the only species of bee native to the United States that is reared commercially. In the Northeast, commercial B. impatiens colonies are used by farmers to enhance pollination of greenhouse tomatoes as well as a number of other crops, including blueberries and pumpkin . As commercial colonies are also available to researchers, B. impatiens has been used in research targeting diapause , allowing us to compare field based studies to lab studies using the same species.In the fall of 2019 and 2020, we performed systematic searches around the entrances of B. impatiens nests to look for evidence of overwintering queens . We searched for evidence of overwintering queens around four nest sites: three nests were located in 2019 during other field work and one nest was located in 2020 during a meandering search for bumblebee nests at Appleton Farms and Grassrides. Although we did not confirm that these queens were genetically related, we assume that each aggregation largely consisted of siblings. Although we have observed queens excavating hibernacula around nest entrances in woodlands, forests, and grasslands , we did not search for overwintering queens around nest sites in grasslands because it was far more difficult to find evidence of hibernacula under dense grassland vegetation than in leaf litter. From the first week of September until the end of October, we visited nest sites once per week to mark the locations of overwintering queens . During each visit, we searched 7–9 quadrats around each nest site for evidence of B. impatiens hibernacula . Quadrats were laid out at varying distances from nest site entrances, with the furthest 10 m from the colony entrance . The same quadrats were searched each week for approximately 10 min per visit. All potential overwintering sites were marked with uniquely numbered aluminum tags.To confirm the presence of overwintering queens in the soil, we excavated 141 marked overwintering sites in late November and early December across the four nesting aggregations; approximately 35 overwintering sites, selected randomly, were excavated from each aggregation . To excavate queens, we gently dug around the marked entrance of overwintering sites. Many queens were found just a few centimeters below the surface, in distinct hibernacula . However, if queens were not immediately located, we dug approximately 10 cm below the soil surface before ceasing our search. Differentiating between live and dead queens was fairly straightforward. Live queens had pristine coats and became active upon disturbance, whereas dead queens were often partially decayed or covered in mold . We measured the depth at which each queen had been buried, and gently dislodged queens from their overwintering positions using either a bristled brush or a digging utensil. Live queens were immediately placed on ice to prevent complete disruption of overwintering. We measured the intertegular span of all live queens using digital calipers, and marked each queen with a uniquely numbered, fluorescent plastic tag , after which we reburied queens in their original positions and replaced the numbered metal plant tag in the soil above each queen. In order to ensure bees were collected before they normally emerge, we re-excavated queens in late March , 3–4 weeks before B. impatiens queens are normally observed nest-searching in this study system.