Month: April 2025

We used a one-sample t-test to test the null hypothesis that the pollination efficiency of honey bees equals the efficiency of the average non-honey bee floral visitor . Since honey bee relative efficiency did not differ between agricultural and wild plant species , data from all plant species were combined. The best multiple regression model selected from a set of candidate models of environmental variables revealed that the network-level frequency of visits by honey bees is positively related to the first principal component of temperature bioclimatic variables , where higher values correspond with higher overall temperature, higher isothermality, lower annual range and lower seasonality . Honey bees were also more frequent floral visitors in mainland networks compared to island networks . Perhaps surprisingly, our regression model revealed no effect of the honey bee’s native status on honey bee numerical importance . Release from pathogens and parasites often contributes to the success of introduced species ; this factor may be unimportant in honey bees because many of their pathogens have spread worldwide due to trafficking of domestic colonies . Nevertheless, it is noteworthy that eight of the ten networks with the highest relative frequency of honey bee visits come from introduced range localities, and that in five of these networks, honey bees accounted for more than half of the total visits recorded . While Abe et al. found that honey bee dominance in the Ogasarwara satellite islands was driven by an introduced lizard’s preferential predation on native pollinators, black plastic garden pots further studies are needed to understand why honey bees reach high abundance in some parts of their introduced range, but not others.

Also surprising is our finding that study year was unrelated to honey bee numerical importance , given the high mortality in managed honey bee colonies reported over the last two decades . Agents responsible for increased mortality in managed colonies can also affect wild or feral honey bee colonies , but ongoing research also reveals the resilience of unmanaged honey bee populations to mortality agents such as parasites and pathogens . In our pollination networks, the degree to which honey bee individuals are coming from managed versus unmanaged colonies likely varies based on geographical location and proximity of the study site to agriculture. However, in one network with high honey bee numerical importance , genetic testing indicated that the majority of the honey bee foragers were derived from feral, Africanized hives . Although honey bees are numerically dominant pollinators in many networks, their importance as floral visitors to individual plant species varies widely. An examination of 46 pollination networks that provide data on each studied plant species yielded 1629 plant taxa within these networks. While some plant taxa species are found in more than one network, we treat each plant species within each network independently because our goal is to examine the frequency with which honey bees visit each plant species within discrete communities. Across these 1629 plant taxa, we found a strongly, positively skewed distribution of honey bee visitation frequency . Honey bees were the only documented visitors of 5.34% of plant taxa , and contributed the majority of visits to 15.16% of plant taxa . However, honey bees also failed to visit the majority of plant taxa .

Restricting the analysis to plant taxa with ³ 10 visits recorded to minimize extreme values due to low sample size did not qualitatively affect our results . In this data subset, honey bees were the sole documented visitors of 3.44% of plant taxa , contributed the majority of visits to 17.84% of plant taxa , and failed to visit 50.31% of plant taxa . Our finding that honey bees numerically dominated a number of plant taxa is perhaps unsurprising given their ability to recruit nest mates to spatially and temporally abundant floral resources . However, it is noteworthy that this pattern holds true in their introduced range, where floral resources monopolized by honey bees presumably coevolved with native pollinators. This analysis cannot distinguish whether honey bees dominate certain floral resources because they displace other pollinators or because they have the ability to profit from floral resources not valued by other pollinators. However, the data do suggest that honey bees possess the potential to disrupt interactions between plants and other pollinators in the majority of natural communities in which they occur. On the other hand, our finding that honey bees are frequent floral visitors to only a small subset of the plant species in a community is consistent with studies investigating honey bee colony-level resource use and underscores the importance of maintaining robust, diverse communities of non-honey bee pollinators for the persistence of the majority of plant species in natural communities. While our analyses of pollination networks worldwide reveal that honey bees are exceptionally abundant and generalized floral visitors, our analysis of pollination efficiency of honey bees reveals that they are average pollinators with respect to their pollination efficiency . Using a dataset of 35 plant species spanning 23 plant families that exhibit a diversity of flower sizes, shapes, and colors, we compared honey bees and non-honey bee floral visitors with respect to seed set, fruit set, or pollen deposition resulting from single floral visits . The relative pollination efficiency of honey bees did not differ between the 16 agricultural and 19 non-agricultural plant species , perhaps because flowers of agricultural species , squash , tomato often closely resemble those of their wild relatives. Overall, we found no evidence that the pollination efficiency of honey bees consistently differs from that of the average of the non-honey bee floral visitors considered in these studies . Since the importance of a particular pollinator to a given plant species is often calculated as its per-visit efficiency multiplied by its visitation frequency , it seems reasonable, given our results, to assume that the ecological importance of honey bees as pollinators in any community is satisfactorily estimated by their visitation frequency. However, since honey bees are known to exhibit poor efficiency at pollinating certain plant taxa , we caution that careful studies are needed to demonstrate the importance of honey bees as pollinators to particular plant species. Further, in at least one case, high visitation frequency by a pollinator damaged raspberry flowers and led to reduced reproductive success . On plant species and in plant communities where honey bees reach high visitation rates, a similar negative relationship between visitation frequency and plant reproductive fitness may occur and is worthy of investigation . As a numerically dominant, super-generalist pollinator, honey bees may influence the fitness and behavior of competing pollinators, enhance as well as reduce plant fitness, and facilitate the spread of non-native weeds and pathogens . Given the ecological importance of honey bees, there is little doubt that changes in their distribution and abundance will impact the evolutionary trajectory of co-occurring mutualists and competitors, and likely the long-term eco-evolutionary dynamics of communities in which they take part. Our results underscore the urgent need for more data on how honey bees, and the potential loss thereof, shape the ecology and evolution of plant-pollinator interactions on global and local scales.Drosophila suzukii Matsumura is an economic pest of small and stone fruit in major production areas including North America, Asia and Europe. Female D. suzukii oviposit into suitable ripening fruits using a serrated ovipositor. This is unique compared to other drosophilids, square plastic plant pots including the common fruit fly, D. melanogaster, which oviposit into overripe or previously damaged fruit. Developing fruit fly larvae render infested fruit unmarketable for fresh consumption and may reduce processed fruit quality and cause downgrading or rejection at processing facilities. In Western US production areas, D. suzukii damage may cause up to $500 million in annual losses assuming 30% damage levels, and $207 million in Eastern US production regions.

Worldwide, the potential economic impacts of this pest are staggering. Pesticide applications have been the primary control tactic against D. suzukii both in North America and in Europe. The most effective materials are those that target gravid females, including pyrethoids, carbamates, and spinosyns. These applications are timed to prevent oviposition in susceptible ripening host crops. In the Pacific Northwest, many growers have adopted scheduled spray intervals of 4–7 days. This prophylactic use of insecticide is unsustainable as growers have a limited selection of products and modes of action. This could ultimately lead to D. suzukii becoming resistant and may cause secondary pest problems because of negative effects on beneficial organisms. Furthermore, production costs have increased substantially in crops where D. suzukii must be managed. Effective sampling methodology for D. suzukii is lacking despite extensive efforts to improve trap technology or determine effective fruit infestation sampling protocols. Theoretically, traps to capture adult flies should aid growers in the timing of spray applications so that insecticides could be used more judiciously. Traps baited with apple cider vinegar or a combination of sugar-water and yeast are currently used to monitor adult D. suzukii flight patterns. However, without standard methods for trapping or management thresholds based on trap count data, it is questionable how much is gained by establishing and monitoring traps in crops. Establishing, monitoring, and maintaining traps is very labor intensive and the costs do not justify the benefits for many growers. Historically, trap data has not provided a reliable warning against D. suzukii attack, especially for susceptible crops in high-density population areas where considerable oviposition can occur in short time periods. Currently, no significant differences are found in any traps used for monitoring D. suzukii given differences between crops and environments where traps have been tested. Monitoring fruit infestation levels to guide management may also be impractical. It is unclear how many samples would be needed to accurately determine infestation levels. Furthermore, by the time larvae are detected in the fruit, it is too late for management action and damage has already occurred. No detailed studies could be found using monitoring for fruit infestation for this pest, and precision of sampling methodology is currently unavailable. Degree-day , or phenology models, are standard tools for integrated pest management in temperate regions and are used to predict the life stages of pests in order to time management activities and increase the effectiveness of control measures. Degree-day models work best for pests with a high level of synchronicity and few generations. Our data suggest that D. suzukii has short generation times, high reproductive levels, and high generational overlap compared to other dipteran fruit pests. Given this life history, stage-specific population models represent an alternative and potentially more applicable tool for modeling pest pressure. Pest population estimates may be greatly improved by employing additional tools such as mark recapture and analytical or individual-based models. The ability to describe and forecast damaging pest populations is highly advantageous for fruit producers, policy makers, and stakeholder groups. Many such studies have been directed at forecasting populations of medically important insect species. The major factors affecting survival, fecundity and population dynamics of drosophilids include temperature, humidity, and the availability of essential food resources. Therefore, an improved understanding of the role of temperature on D. suzukii may provide for a better understanding its seasonal population dynamics. In this paper, we present a population model for D. suzukii that represents a novel modification of the classic Leslie projection matrix, which has proven to be one of the most useful age structured population models in ecology, with applications for diverse organisms including plants, animals, and diseases. Our modification accounts for the effect of temperature on the survival and fertility of D. suzukii in calculating population growth of the organism. Typically, researchers have introduced elements of environmental stochasticity to matrix models to study environmental effects on population trajectories. However, our approach relies on temperature-dependent estimations of age specific fecundities and survival that are determined by models fit to life table data generated for multiple temperatures. Our environmentally dependent matrix model is unique in that it does not rely on simulation of environmental effects on populations, but the matrix itself is recalculated at each iteration in direct response to environmental input. Model predictions were run under environmental conditions from different regions to illustrate variation between and within study sites in different years. These simulations make important predictions about age structure and population trends that have implications for pest management both in a broad sense and with regional specificity.

Accounting for abundance makes multivariate dispersion less sensitive to rare species, which often make up a large fraction of the total species richness in bee assemblages but may contribute little to the pollination services rendered to plants . For these reasons, multivariate dispersion is superior to the traditional approach of using multiplicative or additive partitioning for investigating bee temporal beta diversity with respect to characterizing individual-level bee assemblage composition, as well as temporal turnovers in ecosystem function. To calculate multivariate dispersion, we performed a non-metric multidimensional scaling ordination based on a dissimilarity matrix of abundance-weighted bee assemblages in all possible pairs of samples across all plots . From this ordination, we calculated the multidimensional centroid of the samples from each plot, and then computed the mean distance between each plot’s centroid and its constituent samples. The resulting dispersion score for each plot thus measures the degree to which the species composition of each plot’s bee assemblage turns over through time. Dispersion scores of reserve and fragment plots were then compared using Welch’s two sample t-tests. As with our analyses of temporal gamma and alpha diversity, we repeated all analyses with the temporal beta diversity of native plants as an added independent variable . Models that included plant temporal beta diversity yielded poorer AICc scores in all cases; thus, we did not include plant temporal beta diversity in our final models.Across our two years of sampling, we found consistent differences in bee assemblages occurring in reserves and fragments, despite the known tendency for bee faunas to exhibit considerable inter-annual variation at a given locality . Compared to reserves, big plastic pots fragments harbored bee assemblages that were less diverse with respect to all three components of temporal diversity .

While all metrics of bee diversity and abundance varied with time, differences in bee diversity between reserves and fragments were remarkably constant through time . Individually scrutinizing the three components of temporal diversity allowed for a high resolution characterization of the temporal structure of bee assemblages in intact and fragmented habitats; these analyses also yielded further insights into the potential consequences of bee diversity loss for ecosystem function in fragmented habitats in our system. Reduced species richness is one of the most commonly reported effects of habitat fragmentation on bee assemblages . Though our reserve and fragment plots did not differ systematically with respect to the composition of floral resources , it is possible that decreased availability of nest sites within foraging distance of key host plants or increased vulnerability to demographic stochasticity due to isolation or small population size may have contributed to reduced bee species richness in fragments. Analyses of the temporal gamma and temporal alpha components of bee species richness yielded qualitatively similar results; however, the impact on each of the two temporal diversity components may have distinct implications for the conservation of bees andecosystem function. The temporal gamma component of bee richness provides information on the habitat conditions and locations that support the greatest total number of bee species or species of particular conservation concern; as such, it is the most useful metric for developing conservation strategies aimed at bees. On the other hand, the pollination effectiveness of a particular bee species for a particular plant species may depend upon the timing during which the interaction between bees and plants takes place or upon the bee species’ functional complementarity with other, temporally co-occurring pollinator species .

Detecting potential impacts of climate change on the phenological matching between bee species and the plants they pollinate also requires examining the composition of bee assemblages at discrete points in time. Thus, in the face of a changing climate, effective strategies aimed at conserving bees and the ecosystem function they perform should account for both the temporal alpha and gamma components of bee richness. As with patterns of bee species richness, patterns in the temporal gamma and alpha components of bee assemblage evenness are in qualitative agreement with each other. Assemblage evenness is an important driver of ecosystem function , including pollination , but remains an under-appreciated aspect of pollinator assemblage dynamics . Reductions in the temporal alpha component of bee assemblage evenness in fragments may result in decreased frequencies of interspecific encounters among bee species; such encounters have been shown to enhance pollination efficiency via altering bee foraging behavior . On the other hand, reductions in the temporal gamma component of bee assemblage evenness may result in a stronger reliance by plant assemblages on a small subset of numerically dominant bee species, and consequently, reduced stability of pollination services . In contrast to patterns of bee species richness and assemblage evenness, overall bee abundance did not differ between reserves and fragments. This pattern was caused by reserves having higher bee abundance in spring and fragments having higher bee abundance in summer . This treatment-by-sample interaction appears to be driven by the higher relative abundance of generalist bees in fragments ; many generalist species in our system reach peak abundance between late June and August. Generalist bees may be more tolerant of habitat fragmentation compared to specialists and have been hypothesized to replace the ecosystem function formerly performed by extirpated specialists .

However, even though generalists in our study numerically compensated for absent specialists when considering the temporal gamma component of bee abundance , reduced bee abundance in fragments early in our study period may threaten the pollination of spring blooming plant species. Temporal beta diversity represents another under-appreciated metric in ecology , and reports on the effects of anthropogenic disturbance on intra annual turnover of biological assemblages remain rare . In our system, decreased temporal beta diversity in fragments may explain how modest reductions in the temporal alpha component of species richness and assemblage evenness in fragments translate into more pronounced reductions in the temporal gamma component . More broadly, decreasing seasonal turnover in an assemblage may result in increasing temporal niche overlap among its constituent species , which may in turn decrease the number of distinct temporal niches created by the assemblage. Decreases in the seasonal turnover of bee assemblages may be especially consequential in cases where bee species tend to interact with a set of preferred host plants throughout their activity season even when new plant species begin to bloom as time progresses . If temporal host-switching is likewise rare in our system, reduced bee assemblage turnover in fragments may jeopardize the reproduction of certain plant species that occupy specific temporal niches with respect to pollination . Examining the temporal beta diversity of bee assemblages thus appears crucial for understanding mechanisms underlying the impact of anthropogenic disturbance on pollination services. Novel selection forces in altered habitats often create ecological filters —environmental or biotic processes that determine which species can establish or persist. The strength of ecological filters depends on the form of disturbance, the natural history of organisms in question, and the strength of other forces shaping community assembly such as dispersal, competition, and natural disturbance regimes . Understanding the extent to which ecological filters shape community assembly represents a central goal of community ecology , and can be used to predict the long-term implications of habitat modifications . Assessing the strength of ecological filters is especially important when evaluating the long-term ecological consequences of habitat fragmentation, one of the leading causes of ecosystem change and biodiversity loss worldwide . Many studies have found strong evidence that ecological filtering drives diversity loss in habitat fragments, both at the local scale and landscape scale . On the other hand, community assembly in fragmented landscapes may also be shaped by stochastic colonization and extinction events typical of island biogeography , or by underlying heterogeneity among habitat patches . While habitat fragmentation could reduce local species richness through different combinations of the above mentioned processes, effective conservation practices will need to account for the degree to which ecological filtering drives species loss . One powerful approach to assess the strength of ecological filters in fragmented habitats is to examine functional diversity, since ecological filters, by definition, black plant pots plastic act on functional traits rather than species . Functional diversity is related to taxonomic diversity in complex ways , and the relationship between the two metrics can provide insight into the mechanisms that drive biodiversity loss in fragmented habitats. For instance, when habitat fragmentation results in strong ecological filtering, functional diversity may decline even if species richness and abundance remains little altered, as may be the case when taxa that thrive in fragmented habitats replace those that are extirpated .

On the other hand, if species loss in habitat fragments mainly results from stochastic extinction events associated with small population size and isolation, functional diversity may be relatively unaffected by the loss of taxonomic diversity, especially in systems with sufficient functional redundancy among species .Here, we evaluate the contribution of ecological filters to diversity loss by taking advantage of an extensive survey of native bees in a species rich ecosystem where we have documented profound reductions in bee species richness associated with urbanization-induced habitat fragmentation . Bees are ecologically important pollinators known to exhibit non-random species loss in fragmented habitats, where specialist species appear particularly vulnerable . Ecological filtering of bees may occur in habitat fragments when fragments experience reductions in the diversity or abundance of plant species that serve as food resource for bees, exhibit altered abiotic conditions due to influences from the surrounding matrix , or fail to contain the correct spatiotemporal configuration of food and nesting resources . Habitat fragmentation may also reduce bee diversity via processes not related to ecological filtering; for instance, when the isolation of habitats disrupts dispersal processes crucial in buffering bee populations from year-to-year variation in the local and temporal distribution of floral resources . We assess the strength of ecological filters by addressing four questions. First, to what extent does fragmentation impact bee functional diversity? Loss of functional diversity more severe than that predicted by a null model of random species loss would lend support for the importance of ecological filters . Second, do bee assemblages in fragments exhibit distinct taxonomic or functional compositions compared to those in reserves, as would be expected when ecological filtering causes the assembly of novel communities ? Third, do bee assemblages in fragments exhibit lower taxonomic or functional beta diversity among plots, as would be expected when ecological filters select for or remove common sets of functional traits in altered habitats ? Lastly, are bee assemblages in fragments composed of taxa with larger range sizes relative to those in reserves? Given that range size tends to be positively related to niche breadth , a shift to more cosmopolitan species in fragments is expected if ecological filtering precludes the persistence or colonization of species more specialized to the local ecosystem . Answering these questions will yield insights into the mechanisms that drive bee species loss in our study system as well as provide information on the potential conservation value of scrub habitat fragments .Study system: Field data were collected between April and August of 2011 and 2012 in coastal sage scrub habitat in San Diego County, California, USA, and are detailed in Chapter 1 of this dissertation. We surveyed one-hectare study plots belonging to two categories: large natural reserves , and habitat fragments surrounded by urban development. This is the same system of reserves and fragments previously used to study the ecological effects of urbanization-induced habitat fragmentation . In 2011, we surveyed eight study plots ca. every 2-4 weeks; in 2012, we surveyed 17 study plots ca. every 3-5 weeks. During each survey at each study plot, the first author deployed 30 bowl traps between ca. 0900 h and 1500 h, and collected free-flying wild bees via aerial netting . Concurrently, we also documented the identities of native, insect-pollinated plant species in bloom at each study plot by walking through pre-planned paths that allowed an observer to visually survey the entire study plot . All collected bees were individually mounted and identified to species or morphospecies within genus, hereafter referred to collectively as “species” . This sampling effort resulted in a dataset of 11,037 native bees belonging to 216 species in 52 genera and 6 families, after the exclusion of bee specimens not identifiable beyond genus .

Leaf out and flowering occur earlier now than historically, and plants are moving upwards in elevation or polewards in latitude as the climate warms . These shifts in space and time could have important community and ecosystem consequences, particularly if the species closely associated with plants are not responding synchronously . However, interactions between plants and their associates, and species interactions more generally, are understudied in the context of long‐term anthropogenic environmental change because long‐term data on these interactions are relatively rare . Here, we used pressed plant specimens from the northeastern US to determine how herbivory has changed over the last 112 years. While herbarium specimens have been used to quantify long‐term changes in plant phenology and morphology , they are used much less often to investigate changing interactions between plants and their antagonists. Changes in these antagonistic relationships could have consequences for ecosystem structure and function. For example, insect herbivory is a major driver of plant coexistence , biomass production , nutrient turnover and retention , animal composition , and ecosystem trajectories . Additionally, long‐term herbivory data from herbarium specimens could inform pest management by revealing trends over time, thus allowing us to determine if monitoring of certain crop or forest plants should become more frequent as global change progresses and whether new pest control strategies might be needed. Predictions for how climate warming across seasons and years may affect herbivory are complex , square pots and empirical assessment of herbivory change is limited because there are few existing long‐term observational or experimental datasets .

In experimental settings, insect survival and fecundity increase with warmer temperatures . We might thus predict that herbivory should increase with climate warming unless temperatures exceed thermal optima of the herbivores . However, warming in winter could increase insect mortality by reducing snow pack, which exposes insects to very low temperatures, and/or by disrupting winter diapause . Total herbivore damage may be further complicated by a number of additional interacting factors, including changes in insect predation , plant and insect phenology , and community composition . As a result of these or other mechanisms, patterns of herbivory across latitude and elevation are highly variable . As a consequence, spatial variation in herbivory might not reliably capture how herbivory has changed over time and with recent anthropogenic transformation of habitats and climate. Over the past few decades, it has become clear that, alongside climate change, urbanization may have profound effects on interactions between plants and their insect herbivores. Urbanization is increasing at unprecedented rates , and its effects on herbivore damage to plants remain poorly described. A number of studies show that certain leaf‐feeding insects are excluded from urban habitats, which may reduce herbivory . However, some herbivores may benefit from urbanization if they are able to escape their natural enemies or if the urban heat island effect increases population sizes . There remains, therefore, large uncertainty about whether the relationship between herbivory and urbanization can be generalized, and how climate change and urbanization may interact to influence the effects of insect herbivores on their plant hosts. In the northeastern US, mean yearly temperatures have risen 0.8°C from the early to the late 20th century. Winter temperature, a key determinant of insect herbivore survival , has increased by 0.9°C, more than other seasons .

Some urban areas within the region, such as in the Boston metropolitan area, have expanded rapidly , whereas other parts of the region remain largely rural. Here, we examine trends in herbivore damage over this time span across four plant species with varied life histories and broad geographical distributions. First, we tested whether insect herbivory was associated with warmer temperatures and urbanization. Second, we explored a proximate, biotic driver of changing herbivory over time—insect herbivore presence—with occupancy models constructed from the present‐day insect herbivore observations collected by citizen scientists. Although we do not have matching temporal data on the insect herbivores, if shifts in insect herbivore populations are driving changes in herbivory, we would predict that insect herbivore occupancy should be positively associated with the same suite of climate variables driving herbivory change through time. Effects of urbanization are more difficult to predict and might add to or counteract effects of climate change on herbivory.New England encompasses much of the northeastern USA and includes the states of Maine, Vermont, New Hampshire, Massachusetts, Connecticut, and Rhode Island. The climate isvariable and characterized by a strong latitudinal gradient in temperature and varied geography, including mountains in Maine, Vermont, and New Hampshire. Detailed information on the climatic history of this region is included in Appendix S1 and Figure S1. We sampled herbivore damage on four focal taxa in this region: shagbark hickory [Carya ovata], swamp white oak [Quercus bicolor], showy tick trefoil [Desmodium canadense], and wild lowbush blueberry [Vaccinium angustifolium]. These species are native to the northeastern United States and are subject to natural herbivory in the wild but were identified by local experts as only rarely subject to herbivory within herbaria.

Our focal plant species span a broad range of life histories and are eaten by different insect herbivores, some of which are specialized. Desmodium canadense is an understorey herb that is a host plant for many butterflies and moths. Carya ovata is a common canopy tree that provides food for a variety of wildlife. Vaccinium angustifo‐ lium is a low‐lying shrub used for commercial blueberry production. Quercus bicolor is a canopy tree used as a timber species. These species are located on distant branches of the angiosperm phylogeny, including Fagales , Ericales , and Fabales , and thus represent highly divergent evolutionary histories. The primary insect herbivore species associated with these plants in our study region are listed in Table S1. We quantified herbivory on all specimens collected in New England and preserved within the Harvard University Herbaria, excluding duplicate specimens—those collected on the same day and in the same location as a specimen already sampled—and those without county‐ level location data or full collection dates . In total, we quantified herbivory on 123 C. ovata, 89 Q. bicolor, 149 D. canadense, and 215 V. angustifolium specimens , spanning the years from 1896 to 2008. None of the specimens we examined were type specimens, and the species we included in our study were selected to be common and thus well‐represented in herbaria. We scored herbivory by overlaying a grid of 5 cm by 5 cm cells, and scoring presence or absence of damage in five randomly selected grid cells, ensuring that selected cells had at least one‐ fourth leaf cover . We focused on “chewing”—leaf removal by herbivores with mandibles, likely including Lepidoptera , Coleoptera , and Orthoptera —because it was the most common type of herbivory on specimens and is the subject of most ecological studies on herbivory . When leaves overlapped within a grid cell, we carefully lifted upper leaves to check for herbivory on the leaves below. However, such approaches do not cross over well to herbarium specimens. For example, herbarium specimens frequently have overlapping leaves that are fixed to specimen sheets. These vouchers can be carefully manipulated and inspected for the presence or absence of herbivory—as we did here—but do not allow us to accurately calculate leaf area. We scored herbivory by examining specimens under a microscope, allowing us to document evidence of herbivory even where the percentage of leaf removed was small, square plant pot and to better determine whether leaf damage occurred before the specimen was collected or subsequently in the herbarium . Digitized specimens that might be used for image processing to calculate leaf area are typically not high resolution enough to differentiate between these two types of damage. Our scoring of herbivory using a grid‐based system allows us to derive data on the intensity of herbivory while avoiding challenges posed by measuring leaf area. Achallenge in quantifying herbivory on herbarium specimens is that a suite of herbivores feed on pressed plants after they are collected. Thus, post collection damage must be distinguished from damage caused before a specimen was collected. We differentiated between damage by herbivores within herbaria and damage by herbivores on living specimens by the morphology of the damage. We found in this analysis and in a previous study that live plants generally form toughened, necrotic wounds around herbivore damage, but this is not present on specimens damaged indoors post collection. For examples of outdoor damage to live plants and indoor insect damage to specimens after collection, see Figure S2. Given that collectors aim to collect unblemished specimens, our analysis likely represents a down biased and therefore conservative estimate of total herbivory. Nonetheless, estimates are comparable among species and across years under the assumption that collector bias has remained relatively consistent across years. Shifting collection practices—for example, an increased tendency to include damaged specimens in more recent time periods— could give the appearance of changing herbivory through time.

We therefore consulted with curators at several herbaria, and although anecdotal, they reported no knowledge of systematic changes in collection methods that should affect herbivory on specimens. In addition, we note that these biases are not expected to apply to analyses of temperature and urbanization because collectors are not more likely to collect damaged specimens in warmer/cooler or more/less urbanized locations. To further explore potential for collection bias, we recorded the collector identity for each specimen, and then subset the data to collectors that collected five or more specimens. This yielded a total of 32 collectors. We then tested for an effect of collector on herbivory using a generalized linear model with a logit link function in the lme4 package in R with proportion of grid cells with chewing damage as the response and collector identity as the predictor. To examine relationships between climate and herbivory, we extracted various temperature predictors from the PRISM ClimateGroup, 2004 gridded data, all of which were county‐level means corresponding to each specimen collection date . To broadly represent temperatures during overwintering and spring development of insect herbivores, we extracted mean temperatures from January to March during the year when a specimen was collected. This climate predictor is designed to capture a range of mechanisms, including early season phenology and winter mortality. One potential benefit of higher early spring temperatures for insect fitness is faster development out of young, vulnerable stages wherein insects are more susceptible to natural enemies . To more directly capture effects of winter mortality on insects, we extracted mean temperatures in the three coldest months of the year . To represent the potential effects of summer temperature, including heat wave effects on insect mortality, we extracted mean temperatures of the three warmest months in the current and prior year . We included the prior year because many specimens were collected in spring or early summer, prior to the onset of high summer temperatures in the current year. Finally, to represent overall temperatures experienced by insects and plants throughout the year, we extracted mean annual temperatures. This suite of predictors matches the climatic data available for herbivore occurrences . As a proxy for urban development, we used human population density estimated by county and year with data from the most recent US Census . As census data are collected every 10 years, we matched each specimen to the nearest decadal population estimate.We used HOSTS, a database of Lepidoptera host plants collated from primary literature to identify insect herbivore species associated with our focal plants in the continental USA. We focus here on the order Lepidoptera as they are the most frequently collected and recorded insect taxa and have the most complete data on host associations and occurrence. We extracted occurrence data within New England from the Global Biodiversity Information Facility for each herbivore species. All records were collected between 1990 and 2015 and classified as research‐grade observations from iNaturalist, a crowd‐sourcing platform that sources natural history observations from the public. Species with fewer than 30 records were excluded from subsequent analyses because it is challenging to fit species distribution models with fewer than 30 records . The final dataset included 6,853 records for 69 herbivore species, 47 of which were herbivores of Q. bicolor, 16 of V. angustifolium, five of D. canadense, and one of C. ovata . These included one nonnative species, the gypsy moth, Lymantria dispar, which is invasive and was introduced from Europe to Massachusetts, US in 1868 or 1869 .

The cyanogenic potential is measured as the amount of pre-cyanogenic compounds contained in the plant tissue . The cyanogenic capacity is measured as the amount of cyanide released by a given quantity of plant tissue over a unit of time . The enzymatic activity, which can be variable across species and genotypes, will determine how closely measurements of cyanogenic potential and cyanogenic capacity correspond . The interaction between cyanogenic plants and insect herbivores is complex and maybe be affected by several biotic and abiotic factors . As with other chemical defense mechanisms, specialist insect herbivores may have adaptations that protect them from cyanide produced by their host plant including metabolizing it for use of the nitrogen in protein synthesis or sequestering it for their own defense . Generalist herbivores may avoid cyanide intoxication by balancing their diet with cyanogenic and acyanogenic foods .As a defense mechanism, cyanogenesis can operate in two ways depending on the cyanogenic capacity of a given plant tissue: deterrence or intoxication . Cyanogenesis is not a universally effective deterrent and for some insects cyanide may even act as a phagostimulant . Cyanogenesis seems most effective as a deterrent when the cyanogenic capacity is high, release is rapid, and the insect herbivore is an opportunistic generalist rather than a well-adapted specialist . Below a certain threshold, dependent on the herbivore, square pots cyanogenesis is ineffective as a deterrent . Intoxication by cyanide consumption typically occurs when herbivores consume large amounts of plant material with lower cyanogenic capacity . In this sort of situation, cyanide may be released within the digestive track, causing lethal damage or inhibiting growth of the insect .

Plants with higher cyanogenic capacity may be rejected before an insect can consume a sufficient dose. There is variation in the susceptibility of Lima bean to damage by L. hesperus but this has only been catalogued within a small number of commercial cultivars. The first step of this study is to catalogue this variation within a more diverse panel of accessions. From this research, the University of California Davis Dry Bean Breeding program will be able to select better parents and introduce more diversity into the Lima Bean breeding pipeline. In addition to characterizing the variation in L. hesperus tolerance or resistance, understanding more about the possible mechanisms that contribute to this phenotype will help target selection in breeding. To do this a field study in which multiple varieties were vacuum sampled will be analyzed to demonstrate the choice of L. hesperus in the field when multiple varieties are present. It is unknown if cyanogenesis is an effective defense againstL. hesperus in Lima bean. In combination, these studies illustrate the extent of variation in L. hesperus resistance or tolerance phenotypes in Lima bean as well as determine if cyanogenesis is induced by L. hesperus presence and if there is a negative correlated between cyanide and L. hesperus population growth. The other block was a control not treated with insecticides. In 2019, UC 92 plots were vacuumed to in each block to verify that L. hesperus levels were lower in the insecticide treated plots. In 2021, the vacuum equipment was unavailable, so water traps were used to verify the difference in L. hesperus pressure between blocks. Both fields were drip irrigated and conventionally managed. Each variety was planted in a single row in eight plots of 4.5 meters . The plots were randomized within eight sub-blocks. Field notes on days to flowering, growth habit, seed color, and flower color were taken. Yield and 100-seed-weight measurements were conducted after harvest. Four ofthe lines were excluded from the analysis due to poor germination or photoperiod sensitivity.

In a small trial at the University of California Davis Student Organic Farm, nine Lima bean varieties were planted on June 1, 2017, at 38°32’32.5″N 121°46’01.3″W . The field was flood irrigated and organically managed. Every variety was planted in two plots, one of which was randomly assigned a location within each of two blocks. Each plot was 20 feet long and six 30-inch rows wide. Starting at the time of flowering, July 21, 2017, the middle two rows of every plot were vacuumed each week between 11am and 1pm. Samples for each plot were bagged and then frozen. Insects were then transferred to vials of ethanol and adult L. hesperus in the sample were counted and sexed. Adult L. hesperus are highly mobile and can readily fly between small plots. Nymph counts would therefore have been a better measurement to take. However, nymphs were not counted because many were crushed by the force of the vacuum or were two small to be accurately identified with the available expertise. The middle two rows of each plot were harvested measured for total yield and 100 seed weight. The variety UC Lee was removed from the study due to poor germination rates.Five cultivars of California-adapted Lima bean were selected for this study with the aim of representing the diversity of seed size, growth habit, cyanogenic capacity, and tolerance of L. hesperus . All the varieties had white seed coats as this is the market standard for dry Lima beans produced in California. Prior research found that there is not a correlation between seed coat color and cyanogenesis . Originally, one wild accession was included for comparison but due to photoperiod sensitivity and delayed phenology, it proved infeasible to collect samples from these plants in synchrony with the others.In four greenhouse plantings, plants of each variety were individually germinated from seed in azalea pots with approximately two liters of UC Agronomy potting soil mix. Each pot was placed in its own cage with drip line for water and fertilizer . Each cage was randomly assigned a position in the greenhouse. Three flowers and three young podsof each plant were collected one, two, and three weeks after flowering. Flowers were selected with white petals, indicating that the day of sampling was their first to open. Immature pods were approximately 2cm long . Additionally, succulent mature seed tissue was collected four weeks after flowering were sampled. Mature seed samples consisted of a slice of the bean from the opposite side of the hilum from the micropyle of approximately 200mg that would fit into a 96-well plate collection tube . All samples were frozen at -80°C and later analyzed for cyanogenic capacity using the Feigl-Anger paper assay . While some have critiqued this method for being only semiquantitative, it was selected based on the available resources and practicality for analyzing large numbers of samples .Half of the plants were randomly assigned to a treatment group which had adult L. hesperus added to their bug dorm one week after flowering. In the pilot study, 25 adult L. hesperus were added. The resulting level of herbivory was high and, as a result, the susceptible varieties, UC 92 and UC Lee, had insufficient flowers survive for sampling or pod development. In the subsequent rounds of the experiment a total of only 14 adult L. hesperus were added, square plant pot seven one-week-old adult males and seven one-week-old adult females. This level of herbivory preserved sufficient flowers for sampling and pod development on susceptible plants. The one-week-old adult insects were added after the week one flower and pod samples were collected so that those samples had no interaction with the L. hesperus. All cages had a low level of thrips infestation; however, the greenhouses were not treated with insecticide during the study. All L. hesperus introduced to cages in the experiment were one-week post emergence adults reared in a colony founded by individuals collected from Lima bean and alfalfa fields in the fall of 2019 and 2020 . The colony was maintained at 20°C and 12 h of photo period. Adults were held in 30.5cm cube collapsible cage with a bedding of shredded white printer paper, a water-soaked organic cotton round , hulled sunflower seeds, and fresh organic green beans supplied three times a week. The egg-laid beans were moved to rearing tubs where they emerged in approximately 7-10 days. Nymphs were supplied with a bedding of shredded white printer paper and green beans three times per week and moved to adult cages at the time of emergence.The pilot study, September 2019-December 2020 was conducted in a small greenhouse at the Orchard Park Greenhouse Facility at UC Davis and included only two plants of each variety.

The second round conducted in December 2019-March 2020 was conducted in the same greenhouse. In round two, six plants of each variety were planted and divided equally between the two treatment groups. The third round, March 2020-May 2020, was conducted in a neighboring greenhouse that had been enclosed with black plastic drapes to exclude all-natural light. Artificial lights lit the greenhouse in a 12-hour photo period. This was done in an attempt to include a photo period sensitive wild Lima bean accession in the study but even with this treatment, it was not possible to synchronize its flowering with that of the commercial cultivars. In round three, six plants of each variety were planted and divided equally between the two treatment groups. In the fourth round of the study, January 2021-April 2021, the experiment was planted in a neighboring greenhouse with natural light. In round four, eight plants of each variety were planted and divided equally between the two treatment groups. In each round of the experiment, an equal number of plants for each variety and treatment group were planted. However, due to poor germination, not every plant survived to participate in the study. Across the three plantings there were a total of 16 replicates of each variety divided between two treatment groups – so eight plants per treatment. In all rounds of the studies, cages were randomly assigned positions within the greenhouse using the Microsoft Excel random number generator function. All the flower and pod samples were frozen at -80°C for several months before they were processed with a colorimetric assay with Feigl-Anger paper . Defrosting samples were exposed to Feigl-Anger paper for 0-15 minutes, 15-30 minutes, 30-60 minutes, and 60-90 minutes after being removed from the freezer. The results of the assay were scanned and analyzed using the readplate2 plugin on ImageJ 1.52q . This semiquantitative method provided a measurement of the intensity of blue produced by the interaction of volatilized HCN and the chemical treatment of the Feigl-Anger paper. Since the volatilized HCN had to be synthesized from enzymatic activity in thawing sample tissues, a standard of KCN in NaOH solution, used in other studies, was not considered sufficiently comparable to use in estimating the quantities of cyanide released from tissues during each exposure window.To understand how well L. hesperus survive and reproduce on varieties of Lima bean, populations founded by adult L. hesperus added to cages in the cyanogenic response experiment were collected and analyzed after three weeks. The development of L. hesperus is temperature dependent but takes approximately 4 weeks at 20°C, with this being the approximate temperature of the greenhouses in which this experiment was conducted . Three weeks is therefore not enough time for eggs to be laid and the resulting offspring to develop into adults and so the adults collected were survivors of the initial introduction rather than newly developed adults. The number of surviving adults and nymphs were counted for each cage in the treatment group. In rounds three and four, the nymphal instars were identified to indicate the speed of development for L. hesperus on the various varieties of Lima bean. Analysis All statistical analysis was conducted in R version 4.2.1 . For the final analysis, only the 60–90-minute exposure window was used due to concerns that plate position may have affected the results of earlier windows since samples on the outside of the 96-well plate may have defrosted more quickly than samples in the interior of the 96-well plate . This design flaw was not apparent in the time trial since a smaller number of samples were used and so the thermal mass of the plate was lower. It has been noted as an important lesson in experimental uniformity. Prior to publication, a statistical model including plate position may be tested to analyze results from earlier exposure windows.

The disruption in stoichiometry of highly dosage-sensitive components of macromolecular complexes and pathways, across regulatory, signaling and metabolic networks, can negatively affect fitness or be lethal. Thus, partial to complete dominance of one subgenome over the other subgenome may help resolve genetic incompatibilities. Previous studies of ancient allopolyploids revealed that one subgenome may be dominantly expressed and over millions of years retain a significantly greater number of genes. Subgenome dominance has been observed in many allopolyploids, to varying amounts, but not in all allopolyploids nor in any autopolyploids . Thus, the underlying genetic and/or epigenetic mechanisms driving expression dominance remains poorly understood. Previous studies have shown that densities of transposable elements near genes are predictive of which subgenome is more highly expressed. However, if and how much genetic divergence of the diploid progenitors contributes to subgenome expression dominance has yet been evaluated in allopolyploids and especially in vertebrates. An additional whole genome duplication, termed TGD or 3 R, occurred in the teleosts fish lineage, estimated 225–350 million years ago, at the base of the largest and most diverse group of vertebrates. Some clades including Salmonidae, Cyprinidae and Corydoradinae have undergone their own, independent fourth rounds of polyploidization. Cyprinids, the carp family, Grow bag for blueberry plants contain roughly 600 polyploid species derived from potentially at least thirteen polyploidization events.

The family is delineated into eleven subfamilies, including Cyprininae that consists of eleven tribes, of which seven are largely composed of polyploids, Thus, cyprinids are an ideal model family for investigating subgenome evolution following multiple independent polyploid events within vertebrates. To date, to the best of our knowledge, subgenome-resolved assemblies of only three allopolyploid species from the Cyprinini tribe are publicly available, including the common carp, goldfish, and the hexaploid Prussian carp. Some evidence for subgenome expression dominance was uncovered from the analysis of both the common carp and goldfish genomes. However, no evidence for subgenome dominance at the transcriptome level was observed following the analysis of the hexaploid Prussian carp genome. Comparative genomic analysis of the Prussian carp revealed biased duplicate gene retention of certain genes towards one subgenome. This suggests that the genomes of cyprinine allopolyploid cyprinid fishes may exhibit subgenome dominance to varying levels. In this context, the role of transposable element differences, parental effects and/or genetic divergence of diploid progenitor species contributing to observed subgenome expression dominance remains poorly understood. Therefore, the evaluation of multiple independently derived cyprinine allopolyploids can provide valuable new insights into the underlying mechanisms of subgenome dominance. A robust phylogenomic framework for the subfamily Cyprininae is needed to phylogenetically localize polyploidy events and investigate the underlying genetic mechanisms contributing to subgenome dominance in allopolyploid fishes. However, the maternal and paternal diploid progenitors of known polyploids in this group remain largely unknown.

A recent study tried to address this point within this group using three single-copy nuclear loci, but the phylogenetic history of these three genes may not reflect the true history of species relationships within this subfamily. Phylogenomic analyses based on hundreds of orthologous markers from across the genome should reflect a more accurate evolutionary history of the species and more likely to reveal the diploid progenitors of allopolyploids. In the present study, we thus aim to resolve the phylogenetic relationships among several key Cyprininae species, uncover the polyploid origin of three allopolyploid species, identify the closest extant relatives of their diploid progenitors and investigate subgenome dominance and its genetic basis in the allopolyploids. To accomplish these goals, we assemble de novo high-quality reference genomes of twenty-one cyprinid fishes from across five subfamilies using PacBio HiFi long reads. Furthermore, we generate transcriptome data from several distinct organs to investigate subgenome expression dominance in three allotetraploids. Our study provides new insights into the evolutionary history of Cyprininae, including the identification of maternal and paternal diploid progenitor lineages of three independently formed allopolyploids, the genetic basis of subgenome dominance in these allopolyploids, and new large-scale genomic resources for the community as a foundation for future studies.Whole genomes of 21 cyprinid fishes were sequenced with PacBio CCS reads with an average of 32.34-fold coverage and Illumina paired-end 150 bp reads with an average 66.86- fold coverage, in total yielding 2.24 trillion base pairs of raw read data . These datasets were de novo assembled using Hifiasm, yielding high-quality genomes with an average contig N50 size of 23 Mb . The new assemblies ranged in size from 0.81 to 1.83 Gbp, similar to the estimated genome sizes obtained from k-mer analysis of Illumina reads .

A high percentage of Illumina reads aligned against the assembled contigs and high BUSCO scores , suggesting that the biggest proportion of the genomes was assembled . Previous phylogenetic work using three single-copy nuclear loci suggested that three species Procypris rabaudi , Spinibarbus sinensis and Luciobarbus capito are likely tetraploids. To generate chromosome-level genomes, high-throughput chromosome conformation capture reads, at ~100-fold coverage per haplotype, were obtained and scaffolded for each tetraploid with the ALLHiC algorithm. In total, 94.43%, 97.56% and 98.83% of all bases corresponding to S. sinensis, P. rabaudi and L. capito genomes were assigned to 50 pseudo-molecules after manual curation . Strong contact signals of the Hi-C data for all chromosomes of each genome suggest high quality of chromosome-level scaffolding . Homology-based and RNA sequence-based gene predictions were used to annotate all genomes after masking transposable elements , simple sequence repeats , and tandem repeats. The final annotated gene numbers for the three allopolyploids, P. rabaudi, L. capito and S. sinensis, were 45,857, 43,211 and 49,999 , respectively, which were comparable to those of two famous cyprinid fishes common carp and goldfish. The gene number of the rest eighteen species ranged from 23,658 to 32,381, which are similar to the 24,770 for Onychostoma macrolepis and 27,263 for grass carp. BUSCO analysis was conducted to evaluate the completeness of these annotations, which contain an average of 91.6% complete BUSCO gene sets .The overall TE content in the 21 sequenced species ranged from 40.87% to 59.18% . The most abundant repeat class of all species was DNA transposons , of which TC1/mariner, hAT, and CMC were the three top enriched superfamilies . Long terminal repeats account for an average of 11.09% of the genomes, which is higher than reported for zebrafish. Most of our sequenced fishes contained similar long interspersed nuclear element content with that of zebrafish but fewer short interspersed nuclear elements  than zebrafish .We also observed that the median age of DNA transposon families in our sequenced genomes were typically older than those of both LTR and LINE families , which was also found in the zebrafish.Multiple alignments of orthologous genes between each tetraploid and O. macrolepis successfully identified two subgenomes, each of which included 25 chromosomes . To assign each chromosome to a subgenome, a method similar to SubPhaser, a novel subgenome-phasing algorithm using subgenome-specific k-mers as markers, was applied. The allopolyploid origin of several previously determined allopolyploid plants as well as the common carp and African clawed frog Xenopus laevis was supported using this strategy. Therefore, the presence of repetitive kmers, blueberry grow bag which are exclusively or highly enriched towards one subgenome, were sought for each of the three polyploids. We confirmed that two distinct subgenomes, termed ‘subP’ and ‘subM’ , of each tetraploid could be determined based on a suite of 15-mers with unique distribution patterns along each homoeologous chromosome pair, supporting an allotetraploid origin of these three species . To further verify the polyploid origin , we adapted another strategy that involves analyzing TE types and abundances that has been successfully employed to confirm the polyploid history of the African clawed frog, blueberry, sterlet sturgeon, the goldfish and Prussian carp.

This approach is based on the hypothesis that relics of unique transposon types and abundances specific to the two parental species can be used as markers to partition each chromosome to a particular subgenome in an allopolyploid. Frequency analyses of TEs identified between 8 and 16 transposon types in each polyploid genome that were enriched differentially in the subP and subM . These results collectively support an allopolyploid origin for these three polyploid fishes.To estimate the divergence time of each subgenome, we established one-to-one ortholog gene sets from two putative diploid ancestors and the subP and subM genomes of three allotetraploids and calculated the pairwise synonymous substitutions . The divergence-time of diploid progenitors , served as the upper bound estimate of the polyploid event, and can be deduced based on the Ks age distributions of the orthologous pairs . We found that the two subgenomes of L. capito diverged approximately 7.5 to 13.9 million years ago , which is the most recent dateestimate among the allopolyploids examined in this study . In comparison, the divergence of the P. rabaudi subgenomes is estimated at ~15 to 28 Mya. This estimate is similar to the previous divergence times estimates of the subgenomes of common carp and goldfish. The results from our phylogenetic analyses further confirmed that P. rabaudi, common carp and goldfish likely share a common polyploid event, with subP and subM of each species in monophyletic clades . Lastly, the divergence of the subgenomes of S. sinensis was estimated at 10 to 18.6 Mya . Therefore, these three allopolyploid cases, with varying divergence estimates among subgenomes , provides a suitable framework to examine whether genetic divergence of the diploid progenitors contributes to subgenome expression dominance. Mitochondrial genomes are almost exclusively inherited from maternal progenitors, whereas nuclear protein-coding genes are biparentally inherited. Therefore, a comparison of the mtDNA phylogenetic tree and nuclear gene trees enables the identification of maternal and paternal diploid progenitors for allopolyploids. Our phylogenetic analyses using Triplophysa bleekeri or zebrafish as an outgroup provide strongly supported estimates for species relationships and the monophyly of Cyprininae . Furthermore, these analyses revealed three independent polyploidization events: one shared by P. rabaudi, common carp, and goldfish , one in S. sinensis and one in L. capito , consistent with a previous study. Based on the aforementioned phylogenetic analyses and the mitochondrial tree , the subP and subM of these five species denotes the paternal and maternal subgenome, respectively. These analyses also supported three independent allopolyploid origins. The maternal subM of common carp, goldfish and P. rabaudi is most closely related to Tribe Barbini or Acrossocheilini, and the paternal subP is most closely related to Tribe Labeonini. Similarly, a closely related species of Acrossocheilini could have served as the diploid progenitor of the S. sinensis subM, whereas its subP was the descendent of an ancestral fish much older than Smiliogastrini. The formation of L. capito was probably the result of hybridization of two diploid relatives from Barbini. To further confirm the above conclusion, phylogenetic analyses with the whole-genome alignment of 13 species, the fourfold degenerate sites in 1669 genes and CDS of 1669 individual genes were performed. The topologies of all these trees were congruent with each other . Meanwhile, we also observed the differences between overall consesnus species tree and individual gene trees , implying that these topological conflicts may be as a result of incomplete lineage sorting and introgression.Generally, there are four major evolutionary fates for duplicated genes derived from polyploidy events, including 1. duplicate gene retention due to dosage-balance constraints or selection favoring increased dosage of gene products, 2. gene loss or pseudogenization of one duplicate copy, 3. subfunctionalization, the partitioning of ancestral gene functions among the two duplicate gene copies and 4. neofunctionalization, the evolution of novel gene functions in one or both duplicate gene copies. To investigate the frequency of each fate among ohnologs, we analyzed the expression levels across six tissues for a set of positionally conserved syntenic ohnologs that were present in single copy in the genomes of two diploids and retained in duplicate in all three allotetraploid genomes. We identified 4884 to 5,345 gene pairs that had expression patterns consistent with duplicate retention due to dosage-selection, 226 to 348 due to non-functionalization, 9 to 14 due to subfunctionalization, and 223 to 420 dueto neofunctionalization . Examples of expression divergence consistent with subfunctionalization and neofunctionalization for each allotetraploid are shown in Supplementary Fig. 20. However, we should notice that the low level of subfunctionalization inferred could be due to the relatively small number of tissues examined.

All yields were recorded on a dry matter basis , adjusting plot weights by the average dry matter percentage.An experiment was established in April 2018 consisting of four replications laid out in a randomized complete block design. This trial was a large dormancy evaluation but included all populations in the yield trial described above except for the cycle one phenotypic selection population – NY1221. The trial was located in Davis, CA on a Yolo silty clay loam . Plants were sown in trays in the greenhouse 2 months prior to transplanting. Plots consisted of a single row of 25 plants spaced 30 cm apart with a 90cm gap between plots within rows and 60cm spacing between rows. Plants were harvested throughout the season when they reached the target maturity of bud to early flowering stage using a self-propelled forage harvester. Fertilizer was applied to maintain P and K at appropriate levels for high yielding perennial forages, with weeds, insects and other pests monitored and standard control measures applied if necessary. Plant height was measured 25 days after the final harvest in October in 2018 and 2019. Plant height was considered the distance from the soil surface to the tallest point of the plant at its natural height . A single measurement was taken from each plant and an was average height across all plants in the plot was determined.All analyses were performed using R statistical software . For the yield trial, grow bag gardening analysis of variance was conducted to estimate the effects of population, location, and harvest on forage yield.

Locations were analyzed independently followed by a multi-environment analysis. In addition to an overall annual yield analysis, the yield of individual harvests within location was conducted and across locations, harvests were analyzed by season, with spring , summer , and fall groups included. For the single location models, population, block, harvest, and population × block interaction were treated as fixed effects . For the multi-environment models, population, location, harvest within location, block within location, population × location interaction, and population × harvest within location were treated as fixed effects. Means and standard errors were calculated on a per harvest basis for each population using the emmeans package . Pairwise comparisons were conducted using the multcomp package . The significance threshold was set at 0.05 using the Tukey method for multiple comparisons.Forage yield differed among populations at each location and in the overall analysis . Forage yield also differed among populations for the different seasonal harvests at each location. The base population , phenotypic selection cycle one , genomic selection cycle one – high , and genomic selection cycle one – random were the top performing populations overall across both locations; the genomic selection cycle one – low and genomic selection cycle two populations were the poorest performing. These trends were also observed in the seasonal breakdown of harvests, with the exception of GSC1-L which performed well in the autumn harvests. In Ithaca, NY0847, GSC1-H, and GSC1-R had the highest yield per harvest over the four-year duration of the trial. GSC1-L had low yield in the spring harvests and overall. GSC2-H and GSC2-L were consistently the lowest yielding populations.

Although NY1221 was not significantly different to top performing varieties in the seasonal breakdown of harvests, the GSC1-H population had a significantly higher forage yield than NY1221 overall. In Tulelake, there was less separation between populations. NY0847 and GSC1-H yielded significantly more biomass than GSC1-L in spring, summer and overall; however GSC1-L outperformed GSC1-H in the fall. There was no significant difference between genomic selection and phenotypic selection in Tulelake.Fall height was relatively consistent among populations in the Davis experiment . The only significant difference was that GSC1-L had taller autumn regrowth than GSC2-H. A wide range of broad sense H2 was estimated for forage yield across all harvests . H2 for total annual forage dry matter yield over years was 0.93 at Ithaca and 0.57 at Tulelake. Overall H2 across all harvests and locations was 0.53.Dry matter yield is the most important trait for profitable alfalfa production, yet somewhat inexplicably, over the past 30 years, there has been no improvement in on-farm alfalfa yields in the USA . Genomic selection has been shown to increase the rate of genetic gain in many of the major crops grown in the United States, including alternatives to alfalfa, such as maize , in livestock rations . In this experiment, we evaluated and compared populations developed through traditional recurrent phenotypic selection and genomic selection to investigate whether genomic selection could be a viable option to address the lack of yield improvement in alfalfa. In this experiment heritability estimates for DMY are higher than have previously been reported , probably due to the use of ten replications within each location of the trial to obtain reliable estimates of forage yield in a densely sown sward. H2 estimates for Tulelake were lower than Ithaca, due in part to the inclusion of establishment year harvests .

Considering only the populations developed using the genomic prediction model, the GSC1- H population had higher yield than the GSC1-L population, with the population whose parents were chosen randomly falling intermediate between the others. Thus, the model had the ability to shift populations in the expected directions for biomass yield. Across all entries, the GSC1-H population was among the top yielding populations, and GSC1-L was among the lowest yielding. However, the genomic prediction model appeared to break down on the second cycle of selection, with both the high and low GSC2 populations performing poorly. This is concerning, as one of the major benefits of genomic selection is the potential ability to conduct multiple cycles of GS in the span of a phenotypic selection cycle. However, if the model breaks down after a single cycle, this benefit cannot be realized. Nevertheless, conducting a single cycle of GS in the space of a year is still considerably faster than a PS cycle. Approximately 9000 markers were used for the first cycle og genomic selection and fewer were used for the second cycle, which may explain some of the poor performance of the C2 populations; alternatively, the relative value of marker loci could have shifted following the first cycle, so that the model is simply not useful. In addition, there may have been an inadvertent shift in the dormancy of the C2 populations, which could contribute to lower total DMY. The base population and all populations developed through genomic selection were included in a separate trial investigating the autumn height of various alfalfa populations, a proxy for autumn dormancy. The GSC2-H population was significantly shorter than the GSC1-L population indicating selection for plants containing alleles for less fall growth. Future applications of genomic selection should include selection criteria to ensure fall dormancy remains unchanged during the selection process. This also shows a potential risk of genomic selection in a breeding program – the possibility for undesired shifts of non-target traits. Breeders should be aware of this when making selection decisions, and this result highlights the need to measure other traits of importance during the breeding process. The GS model was developed based solely on phenotypic information from plants grown in Ithaca, plastic grow bag and not surprisingly, the selected populations performed better relative to other entries in Ithaca than in Tulelake. This suggests that any potential gains derived from genomic selection require the inclusion of phenotypic information from the target environment. This observation has potentially significant implications for the viability of incorporating genomic selection into alfalfa breeding.

Already suffering from a paucity of breeders and breeding resources, expanding breeding trials to include more environments may not be possible for many breeding programs. Further investigation is required to determine the requisite number of environments that need to be evaluated in order for GS to work robustly. The lack of separation between the base population and populations selected for high yield, either by PS or GS, in the overall analysis provides some insight into what breeders have experienced over the past 30 years of alfalfa improvement. The selected populations have not increased yield in an experiment designed to replicate a commercial production environment. Notably, however, in Ithaca the GSC1-H population performed better than the PSC1 population that was selected through phenotypic evaluation, even though both relied on yield information from Ithaca in making selections or in developing the GS model. Regardless, the lack of DMY gain from the base population using either phenotypic or genomic selection remains a significant concern. Further improvements to the predictive model are possible and may yet result in real gains at the commercial production level. The GS populations evaluated in this experiment derived from a predictive model developed using clonally replicated space-plant yield. A poor correlation between individual space-plant yield and DMY of a densely planted sward is often obtained , so an alternative approach evaluating the DMY of families in densely planted small plots might be a better approach . These families can be bulk genotyped to obtain allele frequency marker data rather than individual genotyping calls . This method better captures commercial yield in the model so more accurate predictions can be made, aligns well with current the current breeding methods in alfalfa, and can be implemented alongside family-based recurrent selection. Genomic selection is still in its infancy in alfalfa; however, our data indicate there is the potential for greater genetic gain with GS than has been obtained with the use of phenotypic selection alone. Significantly more research is required to investigate alternative models and selection strategies across the wide range of environments in which alfalfa is grown. With the cost of genotyping decreasing, new high throughput technologies being developed, and a greater understanding of the alfalfa genome, the potential for GS to improve yield is quite high. The results of this work will be beneficial not only to alfalfa production but also will help guide decision making for breeding of other outcrossing perennial forages.Alfalfa is one of the most important perennial forage crops in the world. It is the third most valuable field crop in the United States in which California leads the nation for hay and seed production, generating in excess of $1B in 2022 . Its high yield and nutritional value are key drivers for California’s dairy industry, the state’s top valued agricultural commodity . In addition to its economic value and importance as a forage, alfalfa provides a host of beneficial ecosystem services. Its nitrogen fixing capabilities and perennial nature promote sustainable cropping systems and contribute to nutrient cycling . Alfalfa also plays a role as an important in sectary and habitat for native fauna . Cultivated alfalfa is predominantly derived from the subsp. sativa, an allogamous autotetraploid . Alfalfa breeding programs are based on recurrent phenotypic selection, with or without progeny testing . They are designed to increase the frequency of desirable alleles in a population while maintaining genetic variability for continued genetic improvement . Breeding goals in alfalfa are characteristic of those in other crops: increasing yield, enhancing forage quality, and improving tolerance to biotic and abiotic stresses . Simply inherited traits with high heritability have been greatly improved through traditional breeding methods; however, improvement in complex, quantitatively inherited traits have been less successful , most notably yield for which there has been little to no improvement over the last 30 years . Long breeding cycles , multiple harvests per year, limited breeding resources, inability to make gains in the harvest index, significant genotype by environment interaction , and selection based on vigor of spaced plants or short family rows are all factors contributing to the low rate of yield progress . Yield improvements in alfalfa in the past can be mainly attributed to improvement of ‘defensive’ traits i.e., improvements in pest and disease resistance . This helps alfalfa reach its yield potential, but it does not result in an increase in yield per se. To select on yield per se, selections could be based on yield data from the first full year of production before plant mortality becomes an influencing factor impacting yield, while persistence could be evaluated at the end of a multi-year trial. Marker-assisted selection is a modern tool that has great potential in addressing the lack of genetic gain in alfalfa yield . The availability of a large number of single nucleotide polymorphism markers, cost effective genotyping assays, and the recent availability of chromosome-scale, haplotype-phased genome assemblies facilitate the dissection of complex traits and provide a pathway for genetic improvement .

There is much variability within land types, even from parcel to parcel or land use categories such as single-family home, residential, high density residential, commercial and industrial. In areas which are highly hostile to bees, pollination is decreased, even among European honey bees, Urban landscape types: Classifications of different landscape cover types were done in somewhat coarse aggregates. For example, park vegetation, road edges, lawns, etc. This was a common choice among students who could quickly select green on maps via Photoshop . Alternatively, some students took the time to trace vegetation with Illustrator or In Design with similar results. Finally, some students also utilized ArcGIS land cover categorizations in GIS data using ArcGIS . Future studies could be done to compare each technique.Landscape design for bees is explored with help of student designers from California Polytechnic’s Landscape Architecture program . Students in both a studio and specialty interest course were given design tasks by author KC over two quarters. Their illustrations and work help to exemplify KC’s vision of designing bee habitat in a thorough ecological, but also provocative and engaging ways. KC’s design ideas for bees are rooted in scientific knowledge and aim to tackle bee pollinator conservation as a multiprong approach. Designs focus around the biological and ecological aspects of bees. The best designs look to celebrate what is unique or interesting about each focal bee. In addition to habitat creation, square black flower bucket wholesale maximization, and conservation, landscape designers can help to show how bees’ stories can be shared.

A seed library network has potential to provide opportunities for improved neighborhood pollinator habitat. Moreover, seed library patrons would be empowered to make positive changes within their vicinities with minimal physical labor and intrusion into private land spaces which are otherwise often inaccessible. Participation in seed library usage would be entirely voluntary for users. A network of seed libraries will act as a system of structural resiliency for urban pollinators. By using geographic analysis, mapping techniques could help to shed light into where seeds are being planted and also where important pollinator plants exist. With the help of citizen science data, areas of low pollinator plantings can be targeted for future landscape design for pollinators.Students were asked to show site visitors about special bees. Figure 1 shows how one student envisioned providing nesting habitat for Megachile bees in a sculptural way, conveying meaning to site visitors . This clever solution helps people to gain landscape literacy about these fascinating cavity nesting bees. Mutualism exists here, presenting opportunity for both bees and humans. Design mutualism is an opportunity for multiple species to benefit from a landscape change . In this case, bees benefit from habitat design for nesting and foraging, while people benefit from gaining landscape literacy about the pollination world around them. Another wonderful design focused on endangered Hylaeus bees in Hawaii. As it turns out the student lives on Oahu, the same island where the endangered Hylaeus species are found. By researching the foraging preferences and last sighting locations of these rare bees, a plan was made to help both conserve and celebrate these now rare bees. Interestingly, the possible conservation area overlapped with an already existing botanical arboretum.

Adding an installation to highlight the special traits of endangered Hylaeus appears to be an opportunity for public education. The student was able to research the face patterns of the local Hylaeus bees and designed an interactive walking tour which would appeal to a large age range of visitors. This project was serendipitous, and holds potential for implementation.Students were challenged to envision and demonstrate what a vegetated landscape looks like from a bee’s point of view, focusing on the valued elements. This project forces students to look at the landscape from their organism’s value system, which is an essential part of good ecological design. Figure 2 shows how Xylocopa favors some forage plants over others in this residential landscape. This student shows clearly which plants have ecological value to Xylocopa with the use of color, in contrast to the colorless portions of the image. This image is particularly good at transmitting meaning to human viewers, helping people to understand this organism’s preferences and landscape opportunities or limitations. Another intelligent “see like a bee” design solution shaped the ear pieces of glasses to look like tubular Megachile nests. The work was completed with an annotated design plan with callouts to highlight favorite foraging plants. The idea of ecologically based ‘bee glasses’ seems like an opportunity for helping capture the imagination of children and with educational presentations. There are many aspects of the bee’s biology, ecology and foraging preferences which could be highlighted and made possibly more memorable with the help of glasses props, for example.An aspect of bee biology which has potential for design is for bees which cavity nest above ground. It is possible for designers to create cavity nesting areas on any vertical surface. The form of these sorts of projects is limitless. One of the best student work’s shows a concept for spelling the desired nesting bee’s genus name . A design like this is fairly simple, yet demonstrates much more knowledge and information than a standard bee box from a standard retailer. Other students looked to maximize wall design space.

One student created a huge silhouette of Megachile and planned drilled holes of the correct diameter all over the entire surface. It is conceivable to imagine that design as both striking and memorable. Other students strove for more abstract geometric patterns, which though artistic in nature, were not effective at communicating as much information about the bees.Author KC has envisioned a new way to help achieve higher quality pollinator habitat in neighborhoods via the installation of free seed libraries. A prototype pollinator seed library was made from a repurposed windowed cabinet and painted to advertise its contents. Since pollinators are suffering from habitat fragmentation and degradation due to human land use activities. Habitat design is critical to solving these connectivity issues today. Improvements to habitat networks are on the forefront of research and design by urban and landscape ecologists. Seed libraries, a grass roots phenomenon, aid in accessibility for people to start their own seeds. These cabinets originated in effort to provide free resource availability and seem like an opportunity for growing pollinator habitat. Seed libraries are small outdoor cabinets which can be curated to a palette of the provider’s choice. The seeds contained within are available free to whoever accesses them. Seed library users are also encouraged to leave seeds for others as well. Seed libraries are a “spin-off” of the popular “Free Little Libraries” program for exchanging books. Cabinet-style libraries are hyper local in design scale, often with one every few blocks in a neighborhood. Designs are often creative, attractive and fun to elicit usage. Specializing seed libraries to help meet the needs of local pollinators has great mutual potential, both for humans, and also pollinators. So far, author KC’s “Free Pollinator Seed Library” has been extremely popular. Well over 800 hundred seed packets have already been exchanged in the months of its existence thus far.The reception of the Grover Beach, California surrounding neighbors and users has been extremely positive. Efforts have been made to create a ‘buzz’ online. Basic information about the project can be found at author KC’s personal website, plastic square flower bucket and it even has its own Facebook page, titled, “Free Pollinator Seed Libraries where author KC can post updates. However, the highest interest occurred by posting on the Nextdoor . Seventy-four people within the immediate neighborhood liked the post and twenty took the time to post comments, all with positive words about the project. Of the commenters, all were within a maximum 12.5 km  radius from the seed library. The average distance of a commenter to the seed library was 6.3 km and the median was 3.4 km and 2.1 mi. Commenters show which neighborhood area they are posting from, which are defined by local’s sense of geographic area, in this containing the following regions: Corbett Canyon, Edge of San Luis Obispo , Fair Oaks-Grand, Grover Heights, Horned Toad Trail, Huasna Valley and Huasna Corridor, Just Off The Pike, Lopez Drive, Oak Park Streets, Oceano, Ocean South, Old Oak Park, S. Oak Park and Trilogy. Most recently a Google Business listing was also made , which has further increased page views and visits. For example, in mid-January 2022 there have been over 1,100 visits to the Google page, which has increased exposure significantly. Therefore, the seed library captured the attention of people in the general geographic area as well as in the local neighborhood. There is a lot of enthusiasm among the neighborhood to help bees. Designing pollinator seed libraries seems to hold a great potential for making the largest positive changes in short amounts of time with limited budgets.

One of the most engaging potential design themes explores the contrasting nature of various bee genera. One student came up with a particularly interesting sculpture idea to celebrate two very different native bees. The student was keen enough to focus on the materials for each bee, wood for Xylocopa, and soil/ceramics for Andrena. The project shows two large bees of contrasting colors diving into the ground with their paths dynamically crossing. This work helps to demonstrate the various nesting substrates each bee would use. Furthermore, the student carefully imagined the body size and shape of each bee. . It would be an impressive sight to see this design implemented. Different bees could be chosen at geographically different places to highlight locally special bees. Other well thought out contrasting bee designs included, showcasing different bee nesting styles or foraging preferences. Some students juxtaposed different style nesters along a human walkway. Others used a human path to separate two very different foraging habitats on each side, for bees with extremely different foraging preferences. Overall, these themes have a lot of potential and should be explored more. Particularly, displaying uniqueness of bees and/or local adaptations seems like an excellent way to support local bee populations.Designing for bees over large areas of human dominated landscapes will require renovation of landscaping with little to no ecological value. Learning to maximize bee habitat with small planting areas is very important. When added together, these small snippets of micro-habitats contribute to pollinator habitat networks, which are essential for resilient landscapes. One of the best submitted images shows one student’s attempt to maximize bee habitat foraging area in their family home. This student wisely recommends more pollinator plantings on the ground level, but also imagines creating more foraging area by utilizing vertical wall space for habitat. Some students were bolder with their designs. They thought through how to maximize the area of foraging plants for bees, whether that was on the roof, walls, driveway, getting rid of grass or paving. Students were urged to think about a design they would like to look at every day, thus, in this way, it was easier for them to imagine if it was their own home or property. Since much of the human-built environment is already in existence, it is very important that we strive to update and augment the ecological functionality of such places for bees and other pollinators.Students tackled the bee habitat map categorization in a variety of ways, each producing effective graphics to demonstrate habitat patchiness of bees in human-dominated environments. Figure 4 shows an attempt at classifying landscape in Oakland, California from Bombus’ perspective. This is a somewhat typical classification of open space, park land, street and sidewalk vegetation. Looking at the landscape with spatial distances between habitat is essential to better understand how bee habitat fragmentation patterns play out for bees at a city-wide scale. Figure 4 was made quickly by using an extension of ArcMap within Illustrator , a more illustrative software. This method was quite quick and accurate, creating quite effective results. Other students tried to streamline the tracing task by utilizing Photoshop’s select by color tool which was perhaps the fastest method, but also lacking in accuracy. More mapping technique results are described below. Other mappers strove to add more detailed information, with varied categories for example, including: natural landscapes, parks, redwood tree dominated areas. This categorization scheme made more sense for the student working out of Mill Valley, California. Some students missed the opportunity to demonstrate human residential areas as possible habitat for native bees. The best projects, also show well the possible geographic connections between denoted patches.

Collected bee specimens were identified to genus both by author KC and with correspondence by bee expert Robbin Thorp , and additionally, in consultation with personnel and comparisons with collections at the UC Davis Bohart Museum . Plant identification was aided by correspondence with former UC Davis Arboretum Director of Horticulture, Ellen Zagory.Data collection was conducted weekly, but compiled into monthly data aggregations to minimize the possibility of sampling omission errors, such as variable detectability . Using this method, it was more likely that a greater magnitude of rarer associations were observed , which is advantageous for a study such as this, seeking to explore the relationships between bees and the plants they utilize. This data compilation method was selected after reviewing bee and other pollinator field research methods as well as much personal trial and error in field and personal correspondence with expert Robbin Thorp. From previous experience we determined that monthly walks produced significantly less association data than compiled weekly walks. Thus, we found that weekly data collection was best for observing ephemeral bee-to-flower foraging associations and monthly aggregations were most effective in understanding bee foraging and flower bloom times . A compiled monthly time step was primarily used for this study as it is a common standard protocol in both bee foraging and plant phenological records, flower harvest buckets such as field guides. As an example, Andrena was seen foraging in the Mary Wattis Brown garden on Ceanothus two weeks in a row.

This association was counted once for the month, not twice, when recording monthly association data. We determined two criteria for measuring a plant’s successful performance, including how many bee genera were attracted to a plant and also, the strength of a bee-to-plant association, with demonstrated repeat foraging events representing stronger associations. Additionally, we sought to determine forage plants utilized by bees which were not included in Table 1. Furthermore, we sought to determine characteristic trends among utilized forage plants, for example, whether they were native or not, and if not, what region of the world they originated from. The analyses were completed in MS Excel in an effort to identify plants missing from the current literature that hold potential for hosting bee foraging and, thus, provide habitat value.Initial investigation into potential pollinator plants revealed that 96 of the 134 of plant genera from Table 1 were included in the Arboretum’s plant collection maps. This indicated that the Arboretum’s records included many of the predictive plants for bees, allowing us to test the majority of the bee-to-plant associations from Table 1. We sought to evaluate to what degree the Arboretum geodatabase plant presence or absence was accurate. We were uncertain about the absolute accuracy of the Arboretum maps, as some new planting projects had taken place since the geodatabase had been completed. We were also interested in studying which weedy plants were used for pollination which were not included on the maps.We used two approaches to assess accuracy of the existing information foraging matrix. Model success is defined here as correct prediction of plants utilized by bees for foraging. First, we looked at how well Table 1 correctly predicted bee foraging overall in aggregate.

This relatively coarse method examines which plants, regardless of bee genus, were successfully both predicted and observed as floral resources. Next, a more precise investigation into the 1:1 association relationships between bees and their forage plants was performed. This statistical testing approach determines the accuracy to which the predictive Table 1 plants were utilized. The existing literature foraging matrix constructed in section 2.2 was validated by compiling field observations made in section 2.5 to determine its efficacy. There are three potential outcomes from this assessment: a correctly predicted presence , an omission error , and a commission error ; however, it should be noted that no correctly predicted absences are possible to assess in this study because the existing literature lists do not designate known absences . This makes traditional assessments of model accuracy using test statistics from a confusion matrix impossible, such as the Kappa statistic .Despite this limitation, it is possible to assess “sensitivity,” also known as the “true positive fraction” from the correctly predicted observations and the “omission rate,” or also known as the “false negative fraction”; these two measures are inversely related and sum to 1 . The third possible outcome is a metric of commission error which assesses the false positive rate. Each of these three metrics will be further described below. In the first case, if a known bee genus is observed in the field that is using a known plant genus this is considered a “correctly predicted” occurrence . For each bee genus this metric is calculated by dividing the count of literature plant genera correctly predicted by the count of all plant genera observed to be used in the Arboretum by that respective bee genus. This yields a “sensitivity score” or true positive fraction. In the second case, if a known bee genus is observed using a plant genus not on the literature list, this is an omission error. The omission rate is calculated by dividing the count of all additional plant genera observed to be used in the Arboretum by the count of all plant genera observed to be used in the Arboretum by that respective bee genus.

This yields an “omission score” or false negative fraction . It should be noted that the sensitivity and omission scores have the same denominator. In the third case, if a known bee genus is not observed to use a known plant genus that is present in the Arboretum, this is a commission error. In other words, the list predicts the bee genus to use the plant, but it is not observed. The commission rate in this study is calculated as a percentage by dividing the count of literature plant genera not observed to be used in the Arboretum by the count of all plant genera from the literature list in common with the Arboretum and multiplying by 100. There are 38 plant genera on the literature list that are not present in the Arboretum and therefore those plant genera are excluded from the error assessment. Finally, to assess the significance or model independence for each bee genus for all observations, a chi-square test was performed on each respective bee genus model result to assess observed versus expected values. This model independence test was conducted using CHISQ.TEST function in Microsoft 365 Excel. This test returns the probability of whether the model could attain the value of the chi-square statistic by chance alone under the assumption of independence. Values for p range from 0-1 and low values of the test statistic indicate independence. The degrees of freedom were calculated by subtracting 1 from the total number of columns used in each respective bee genus model.The completed presence-only bee-to-plant foraging matrix , derived from the literature, contains 23 bee genera and 134 plant genera. Of the 23 bee genera on listed on Table 1, 22 were observed in the Arboretum as well as five additional native bee genera that were not on the list. The only predicted native bee genus not observed in the Arboretum was Colletes, which had a singular association with just one plant genus, Solidago, round flower buckets which is found in Arboretum. In this case most likely either Colletes populations are too disjunct to access the floral resource or there are other lacking resource attributes which prohibited Colletes from using the Arboretum as habitat.The completed observed results of the bee-to-plant foraging matrix contains 27 observed bee genera and 297 observed forage plant genera. Table 2 differs from Table 1 in that results recorded the redundancy of the weekly foraging associations, demonstrating the relative strength of each bee-to-flower association throughout the year. A significant finding of this research is that more than three times the unique bee-to-plant foraging associations were observed than predicted . However, it is clear from Table 2 that plants varied considerably in terms of relative attraction . Appendix 2 shows a complete record of all bee genera predicted versus observed foraging.Observation data were summarized to show the annual pattern of association activity by garden . Bee foraging was well supported by the novel Arboretum plant communities. A full distribution of bee-to-plant associations by garden and month can be seen in Table 3.

Two gardens substantially out-performed all the rest: the Mary Wattis Brown native plant garden and the All-Stars in the Ruth Risdon Storer garden. The plants in each garden supported large numbers of bees, but there were notable differences in function over time. While native plant garden bee foraging peaked in May, the non-native garden peaked in August . Floral resource timing differences accommodate different seasons of bees, who also exhibit staggered emergence and activity months. Additionally, as plants in the native garden often desiccated and rested for the hottest summer months, many of the non-native plants continued to bloom, persisting to provide plentiful floral resources through the hottest months and even fall for summer and fall bees.We examined if Table 1 bee plants in the Arboretum’s map records correctly predicted foraging by bees. As stated in section 2.6, 96 of the 134 predicted plants were included in the Arboretum’s plant record maps. Of the 96 predicted matrix plants which were also in the plant maps 70 were actually used for forage. Wholistically, predicted foraging plant presence was highly correlated with a successful foraging utilization. Within the 84 of 96, or 87.5%, beeto-plant matrix association plants found in the Arboretum plant collections were foraged on by bees, thus, the success rate of the aggregate model indicated high correspondence. The majority of the plants stated to be in the Arboretum geodatabase maps were still present and also used by foraging bees. In total, 84 of the 134 predicted plants, were utilized by bees for foraging. In other words, though 70 predicted plants were also used for foraging and also confirmed on the maps, 14 additional predicted, but unmapped plants , were utilized for bee forage. The majority of Table 1 plants expected for bees were on the maps and 70 out of 84 plants used for forage . This high indication of map accuracy combined with the confirmation of bees foraging on the expected plant list seemed quite promising. Overall, it seemed the habitat relationship model, combined with existing habitat maps, were quite accurate to aid in making predictions in bee foraging habitat use as a whole . Meanwhile, bees were found to forage on many plants not predicted per the Table 1 matrix. Of the “unexpected novel” observed plants, 258 were on the Arboretum maps ,while only 39 forage plants were not on the maps. Interestingly, this is very similar to accuracy percentages of results above . This infers that map records were consistently accurate at providing foraging plant locations and subsequent pollinator association. While the bee-to-plant matrix is predictive of bee foraging the majority of the time, there are a variety of ways to analyze the matrix’s success. The Arboretum mapping accuracy omitted new or weedy plants and therefore some associations seen in Table 2. While these initial results above seem promising, when a more precise analysis is done below, it becomes clear that the individual bee genus models were not as predictive for foraging associations.We analyzed each bee genera by their predicted versus actual foraging data. Each bee genus was compared to the predicted plants it ought to have foraged on versus the observed data. The error analysis matrix and model independence tests presented in Table 4 show the results for each respective bee genus observed in the Arboretum. Error results are reported only for those bee genera listed in Table 1 . Table 4 breaks down the counts relevant to calculating the three aspects of error assessment for this study. The overall average true positive fraction for correctly predicting bee genera in the Arboretum was found to be 0.14 and likewise, the overall omission error rate, or false negative fraction, was found to be 0.86 . The overall average commission error rate for all bee genera was 47.8%, meaning that nearly half of the plant genera reported in the literature that bees are reported to use were not observed to be used in this study.

Such transmission is cited as an ecologically important way for plants to inherit beneficial microbes across generations and for seed associated pathogens to disperse . Vertical transmission has long been observed in grasses, which are hosts to clavicipitaceous fungal endophytes such as Epichloe . Vertical seed transmission has also been observed for non-clavicipitaceous endophytes in Setaria viridis , Triticum , Quercus , and other plants . Floral transmission of microbes into seeds has been studied extensively for pathogens such as Monilinia vaccinii-corymbosi in blueberry and Acidovorax citrulli in watermelon . However, flower-to-seed transmission has also been observed for commensal and beneficial bacteria, for example in Brassica napus . The microbial contributions of the vertical and floral transmission pathways are likely to vary based on a plant species’ pollination mode . Horizontal transmission is the acquisition of seed microbes from the environment, either prior to or after the maturation of the seed as it is still attached to the mother plant or as matured seed disperse and becomes colonized from sources such as air , water , animals , soil , and other seeds in storage . Seed dormancy and germination are likely to represent a very active period of such horizontal transmission, as soil microbes interact with seed exudates and pre-existing microorganisms on and within the seed .The meta community concept was formally described by Leibold et al. , who defined meta communities as sets of local communities that are interconnected by dispersal. This definition arose out of a need to better account for spatio-temporal scales in ecological studies , black plastic plant pots bulk and also included the impacts of dispersal and habitat heterogeneity on community patterns .

Since it was first described, meta community theory has adopted Vellend synthesis that community assembly and composition are driven by four categories of processes: abiotic and host filtering, species interactions, dispersal, and ecological drift . Categories 1 and 2 represent a deterministic or niche-based process of selection where differences in fitness between taxa, species, or guilds lead to differences in their abundances . Finally, drift is the stochastic fluctuation in species abundances, often due to chance birth, death, and migration events . Framing plant microbiomes as meta communities provides an integrated view of the drivers of their composition, function, and evolution, and of the impacts of these drivers on host health . Traditional meta community ecology states that filtering and species interactions occur at the local scale , while dispersal and drift occur at the regional scale . However, categorizing processes as “local” or “regional” is relative to the community that is being studied, and depends on the scales of interest and on defining the boundaries between a local community and a regional meta community. For plant microbiota, including those associated with seeds, the terms “local” and “regional” are contextual because microbes primarily behave at very small scales , although they can be affected by much larger scale factors . Furthermore, microbes can be ubiquitous across habitats at multiple scales, blurring the boundaries between patches of local communities in the landscape of interest . As we apply the first principles of meta community ecology to plant and seed microbiology below, we will therefore use three categories of spatial scale: macro- , meso- , and micro- scales.

Integrating the study of assembly processes across these three scales should give a more complete picture of how microbial communities are assembled, and how emergent community patterns occur at individual scales .Several studies have shown that seed microbial communities differ significantly across geographic locations, i.e., at the macroscale, for example in B. napus , Elymus nutans , Phelipanche ramosa and Pseudotsuga menziesii . For most of these studies, the abiotic factors that are important for structuring these seed microbial communities remain to be identified. However, we can assume that these factors are similar to the ones that drive macro-scale differences in the microbial communities on/in other parts of the plant. In communities associated with leaves, roots, and fruits, such factors include temperature , precipitation , humidity , and soil conditions . In a study of above ground microbial communities in Vitis vinifera, Bokulich et al. found that fungal communities of seeded fruit were associated with net precipitation, relative humidity, and temperature. During dormancy in the soil, the bacterial communities of Noccaea caerulescens seeds were correlated with soil pH and cation composition . Not much is known either about variation in seed microbial community as a function of abiotic factors at the meso-and microscales, although again, much can be learned from studies on other above ground plant tissues. At the meso-scale of an individual plant, microbial communities can vary with tissue location such as canopy height in trees. Unterseher et al. cultured fungi from leaves at different canopy heights in several tree species. They found that species richness was greater in the lower canopy. Harrison et al. went on to use next-generation sequencing in a survey of the needle fungi of Sequoia sempervirens at different height positions, and found that there were distinct communities present at each height across trees.

While they did not measure microclimate variables within the trees sampled, they suggested that the observed variation could be attributed to the amount of sunlight . At the micro-scale , factors such as exposure to ultraviolet radiation and water availability can also be important. Hayes et al. described variation in the bacterial communities and UV radiation along individual flower petals in two sunflower species. They found that while there was no significant difference in community composition along petals, there was variation in UV tolerance in association with source petal position . Another potentially important factor may be water availability, which has been shown to affect bacterial survival, growth, and movement on leaf surfaces . For many macro-scale studies, a major limitation is the use of location as a proxy for environmental conditions, which precludes linking variation in microbial communities to specific environmental factors. Because site effects are impacted by environmental, spatial, and temporal factors, it can be difficult to parse out how location and environment influence seed microbiota . Also, most of these studies do not explore if and how environmental conditions actually select for microbial traits and taxa. In vitro experiments suggest that there is potential for environmental filtering, as demonstrated by thermotolerance in fungal endophytes of desert plants , salt stress tolerance in fungal root endophytes , water stress tolerance in bacterial endophytes , and oxidative stress tolerance in the fungal endophyte Epichloë festucae . Similar characterization of seed microbial tolerance and survival when challenged with different environmental conditions could provide a more mechanistic understanding of abiotic filtering. Such studies would be particularly insightful at the micro-and meso-scales.Variation in plant microbial communities is often studied and interpreted as a result of plant genetics, which represents filtering through host selection. Studies at the macro-and meso-scales have revealed that plant genetics can significantly impact microbial community composition in different parts of the plant, although seeds are clearly underrepresented in the body of literature on this topic. Microbial community variation has been associated with specific genes in leaves and roots of various plants , an approach that has not yet been applied to seeds, as far as we know. Seed line has been weakly associated with microbial community variation in Zea mays and B. napus . Seed accessions of Oryza were also associated with variation in bacterial and fungal community composition, with significant compositional shifts between wild and domesticated accessions . In a study of the bacterial and fungal communities associated with grapes, Singh et al. found that host genotype had an impact particularly within individual sites, procona system whereas abiotic conditions better explained microbial community variation between sites. This is consistent with the notion that host effects are difficult to reveal without carefully controlling for environmental factors, which would suggest, by extension, that environmental factors may have a greater relative impact on seed microbiota than plant genotype. A recent study showed however that the fungal community composition of Quercus petraea internal seed tissue was largely influenced by the mother plant, with only weak significant environmental influences . Studying the roles of plant functional traits in seed microbiome assembly and dynamics provides the mechanistic framework to understand host filtering. Some of the clearest examples of these mechanisms come from the field of plant pathology, where plant traits can be used to predict disease outcomes . One obvious suite of traits to study are plant defenses. As agents of plant regeneration, seeds are one of the most defended plant organs, protected by both chemical and physical defenses .

Some of these defenses come from the mother plant, such as through innate floral defenses in angiosperms . Many studies on plant defense traits are obviously focused on protection against pests and pathogens , but can be extended to other members of the microbial community . A number of studies have been conducted to test how microbes interact with seeds at the micro-scale. Using microscopy, the microbial communities within seeds of Citrullus lanatus and Q. petraea were found to differ in abundance and composition depending on seed sub-structure. Since Q. petraea is a wind-pollinated species, the variation in seed sub-structure colonization observed by Fort et al. suggests physical filtering of microbes during vertical and horizontal transmission. Although few studies have explored the role of micromorphology of developing seeds in microbial community acquisition , there are plenty examples of such micro-scale studies come from work on the floral microbiome. Spinelli et al. used microscopy and fluorescent tagging to study the growth and movement of the bacteria Erwinia amylovora and Pantoea agglomerans on flowers of apple and pear . They found that the bacteria migrate from the stigma to the nectaries along a stylar groove in both species, indicating topographical effects on survival, population growth, and dispersal . Similarly, Steven et al. characterized at the high spatial resolution the floral bacterial communities on apple using next-generation sequencing and found that different flower parts were enriched with different bacterial families . It is intriguing to think that variation in microtopography on flowers and stigmas may contribute to host filtering during the process of flower-to-seed horizontal transmission of microorganisms.The role of species interactions in meta community dynamics is important, but often overlooked in meta community ecology studies . In plant microbiota research in general, much focus has been on pathogen antagonism interactions, for example with an eye toward applications in disease control . However, there is much interest and opportunity to better understand interactions between and among non-pathogens in plant and also seed microbial communities. As with traditional ecology studies, much of the work on species interactions in seed microbial communities focuses on competition and antagonism. For example, Raghavendra et al. inoculated Centaurea stoebe flowers with pairs of fungi and then cultured those fungi out of mature seeds. They always isolated the same single fungus from each pairing out of seeds across parent genotypes, and proposed that competition was the primary driver of selection . Fungi compete for space and resources in Q. petraea seeds , and have negative interactions with bacteria in Populus trichocarpa seeds . Similar competitive exclusion has been observed in floral stigma communities , and in dormant seeds within the soil . However, seed microbes can also coexist via niche partitioning and other interactions. For example, TorresCortés et al. looked at how transmission of several bacterial pathogens impacted the composition of Raphanus sativus seed microbiomes. They found that these pathogens did not alter the composition of the seed microbiome, suggesting that differences in resource usage lead to coexistence between taxa . A more complete understanding of the types and outcomes of microbial species interactions prior to and during seed development is desirable.As with filtering, microbial dispersal to seeds occurs at multiple nested spatial scales, with different mechanisms at play for each spatial scale. For example, at the micro-scale, dispersal from floral stigmas to seeds can be impacted by variation in the level of protection or nutrients that are available to microbial colonizers, which is closely tied to stigma surface topography. The presence of pollen may also be important, as it has been shown that germinating pollen can enhance the flower-to-seed transmission of pathogens and that some bacteria can even induce pollen germination .