Final report for GNE17-146
Project Information
Honeybees (Apis mellifera), which are not native to North America, frequently receive the majority of credit for the economic value of pollination in the United States, largely as a result of the ease with which they can be managed and transported from farm to farm in mobile hives. Several wild native pollinator species provide more effective pollination and provide agricultural crops pollination services valued at $3.07 billion annually. Some crops, such as blueberries, cranberries, tomatoes, peppers, eggplants, and potatoes, tightly hold their pollen in poricidal anthers that require a process called sonication, or buzz pollination, to loosen the pollen grains. Unlike honeybees, Bombus (bumblebees) and bees from such families as Andrenidae (mining bees), and Halictidae (sweat bees) provide very effective buzz pollination. The needs of wild buzz pollinators – especially bumblebees, which constitute 75% of buzz pollinator populations and have annual life cycles requiring early spring forage – differ significantly from those of honeybees, which have perennial life cycles, cannot sonicate, and provide significantly less effective pollination. In order to maximize the health and size of wild native buzz pollinators, our research examined the distinct needs of buzz pollinators as compared to other pollinators in general. To cater to buzz pollinator needs while balancing the needs of farmers who grow crops that benefit from sonication, we interviewed blueberry farmers to structure pollinator habitats that utilize marginal farm land effectively while incorporating the needs of the pollinators. In addition, we conducted common garden trials to determine which native forb species buzz pollinators prefer as forage. The final species of native forbs included species we expected, such as Penstemon digitalis, Monarda fistulosa, and Pycnanthemum muticum. However, we were surprised to see that the aster commonly recommended for pollinators, Symphyotrichum novae-angliae, was the least attractive to buzz pollinators of four species of asters we tested.
1.Consult with farmers on crops requiring buzz pollination to explore:
a.The methods by which they meet their crops’ pollination needs
b.The amount they spend on meeting these pollinator needs
c.The proportion of farmers who depend on on-site pollinator forage and habitat
d.The average amount of marginal land farmers can devote to on-site pollinator forage
e.Those plant species these farmers have found that effectively increase native pollinator populations and increase crop yields
2.Establish relationships with farmers of crops requiring buzz pollination and receive their permission to collect data on native pollinator visitation to their crops. This data will be used to:
a.Compile a list of those species of native pollinators that most frequently visit crops benefiting from buzz pollination
b.Compile a list of perennial and biennial plant species preferred by these crop pollinators for forage when the farmers’ crops are not in bloom.
3.Create a list of native and near-native perennial and biennial plants preferred by pollinators of crops benefiting from buzz pollination using these methods:
a.Interviews with farmers
b.Interviews with native pollinator conservation experts
c.Literature review of scientific and green industry publications
4.Establish common garden plots of herbaceous forbs that will supplement existing display plots of native New England plant species at the University of Connecticut Research Farm in Storrs, Connecticut.
5.Collect data concerning visitation by targeted pollinators to common garden plots by vacuuming insects that visit plots of flowering plants while in bloom and presenting pollen
6.Conduct statistical analysis of pollinator visitation to determine those forage species most preferred by targeted group of native pollinators.
7.Write report that compiles collected data and information.
The purpose of this project is to maximize the health and size of on-site populations of native pollinators for crops that require sonication pollination, such as blueberries, cranberries, tomatoes, eggplants, peppers,and melons. Considering recent declines in pollinator populations and the expense and risk of spreading parasites associated with managed bee populations, tailoring effective on-site pollinator habitats has become important for farmers who grow pollinator-dependent crops. Most literature that provides recommendations for designing on-site pollinator forage and habitats for farmers assume all pollinators share similar foraging needs. While some literature distinguishes itself by supplying profiles of the behaviors and nesting habitats of various pollinators, the authors tend to recommend the same species of flowering plants as forage for all insect pollinators. However, recent research has shown that, if provided enough options, some pollinators strategically will select those plants that provision for particular ratios of protein, lipids, and nutrients. Compared to honey bees (Apis mellifera) foraging in the same landscape, bumble bees (Bombus impatiens) – one of the primary pollinators of sonication pollinated crops – show a tendency to preferentially forage on plant species with pollen that contains higher protein and particular macronutrient ratios. In addition, to avoid or dilute the negative effects of toxic phytochemicals, bumble bees may collect pollen from multiple host-plant species. This research will attempt to develop a selection of plant species and strategies from which farmers growing crops requiring sonication pollination can choose for establishing on-site pollinator habitats that incorporate these new and consequential findings.
Cooperators
- (Researcher)
- (Researcher)
- (Researcher)
- (Researcher)
- (Educator)
- (Educator)
Research
METHODS
- Survey farmers of crops requiring sonication pollination.
Project participants, under the supervision of Dr. Robert Ricard, developed a set of surveys and interview instruments that explored how farmers who grow crops that benefit from buzz pollination meet their crops pollination needs. We constructed a structured survey instrument using Qualtrics consisted of 48 items, including open-ended, opinion-based, and closed-ended questions structured on a five-or seven point scale. The surveys included questions covering the following topics:
- The methods by which they meet their crops’ pollination needs.
- The amount they spend on meeting these pollinator needs
- The proportion of farmers who depend on on-site pollinator forage and habitat
- The average amount of marginal land farmers can devote to on-site pollinator forage
- Those plant species these farmers have found that effectively increase native pollinator populations and increase crop yields
Questions also explored the variety of crops grown, amount of land devoted to crops, and annual yields. From among survey participants, we solicited a select number of farmers to participate in more in-depth in-person interviews. These interviews were used to gain more granular details about the concerns farmers' have involving meeting the pollinator needs of their crops as well as selecting collaborators that would allow for the establishment of on-site pollinator habitats. Surveys were designed to collect quantitative and qualitative data suitable for statistical analysis.
Farmers were found using information from state agricultural departments, state extension offices, the New England Farmers Union, the New England Small Farm Institute, and state Northeast Organic Farming Association.
2. Established relationships with farmers of crops requiring sonication pollination and receive their permission to collect data on native pollinator visitation to their crops.
-
- Record periods of anthesis (bloom time) for each species of crops.
During the first growing season of the study, the anthesis (bloom time) for each blueberry crop species was recorded at each participating farm. We used this data to determine which plant species to include in the neighboring pollinator habitats. It is especially important that, when crops are in bloom and requiring pollination, that the blooms within the habitat do not compete with the crops for pollinator services. Only those species that did not share simultaneous bloom periods were included in habitat plantings.
-
- Collect data concerning pollinator visitation to the farms’ crops.
During the first growing season of the study, white, yellow, and blue plastic trap pans were placed within the plots of highbush blueberries (Vaccinium corymbosum) during before, during, and following anthesis. Each pan was filled with water and dishwashing liquid, which prevented insects from leaving the water because they cannot break the surface tension. Research has shown this technique to be one of the most effective for collecting crop visitation (Campbell & Hanula, 2007; Roulston et al., 2007; Vrdoljak & Samways, 2011). Our own research on willow (Salix) pollinator visitation has proven the effectiveness of this approach.
Bowls were placed in the fields of each participating farm on at least two different days during anthesis. Each time, two-thirds of the bowls were placed at 0830 hours EST on calm days. Half of these two-thirds bowls were collected at 1200 hours EST to determine which pollinators visit early in the day. Another third round of bowls were placed out at 1200 hours EST to determine if some pollinators are later day visitors. All remaining bowls were collected at 1530 hours EST.
Bowls also were placed among fields of crops approximately two weeks preceding and two week following crop anthesis to confirm their presence and abundance correlates to the attraction to the crops’ presentation of rewards.
Samples of pollinators collected using the colored bowl method were stored in 70% ethyl alcohol and subsequently sorted and identified to the lowest taxonomic level using various identification resources, including the insect collections at the University of Connecticut, Yale Peabody Museum of Natural History, and Harvard University Museum of Comparative Zoology, as well as the online key to eastern North American bee species at www.discoverlife.org. Dr. Sam Droege of the United States Geological Survey Patuxent Wildlife Research Center assisted in the confirmation of the final identifications of pollinator species.
3. Created a list of native and near-native perennial and biennial plants believed to be preferred by pollinators of crops requiring sonication pollination with a particular focus on early spring blooming species by using these methods:
- Interviews with experts in the fields of native plants, pollinators, and ecological conservation
- Literature review of scientific and green industry publications
Considering the importance of providing sufficient nutritional sustenance to pollinators, especially bumblebees (Bombus spp.), early in the spring to support their growing colonies, we used a variety of of methods to expand the choices of native plants usually recommended in pollinator forage plant lists. We surveyed several experts in the field of New England native plants we have cultivated in the course of our research. In the course of our search of scientific and industry publications, we paid particular attention to research that focused on pollen provisioning since the protein in pollen appears to be a limiting factor in bumblebee health.
Native plants selected for study in a common garden established at the University of Connecticut Research Farm were evaluated for their relative attractiveness to insects that provide buzz pollination. These plants met the following criteria: (1) native perennial or biennial plants, (2) adapted to agricultural field and woodland border conditions, (3) species representing diverse bloom periods that span from early spring to mid-autumn, (4) species representing a variety of plant families, with varied morphologies and flower colors, and (5) preferably ecotypes that originate from the New England region.
4. Established common garden plots of herbaceous forb species considered native and near-native to the New England region at the University of Connecticut Research Farm in Storrs, Connecticut.
A study site was established at the University of Connecticut Research Farm in Storrs, Connecticut. It consists of 1 meter2 blocks spaced 4 meters apart in a randomized block design consisting of five replicates of each plant species. We included four to nine herbaceous plants within each square meter plot, depending on the growth habit of each species. To reduce weed pressure, we used weed fabric that allows penetration of water and air.
During initial years when we were establishing the common garden plots, we collected preliminary data from display gardens first established at the Research Farm in 2014 and 2015 of approximately 160 plots of perennial and biennial herbaceous plants native to the New England region as part of previous research for the New England Transportation Consortium. Each plot measures 2 meters x 2 meters and consists of 16 to 25 individual plants of one species. There are at least 46 forbs species, 40 grass species, and 20 monocot grass-like species, including 16 sedge (Carex) species. Several of the species have multiple plots, each consisting of ecotypes from different regions from around New England, which we used to observe differences in phenologies and morphologies.
5. Collected data concerning visitation by targeted pollinators by vacuuming pollinators that visit the flowers of plotted plants while in bloom and presenting pollen
Flower visitors were sampled several times weekly from early May until mid-October on calm days between 0830 and 1630 hours EST. Plants were sampled during peak anthesis (full bloom period). We used BioQuip’s Heavy Duty Hand-Held cordless rechargeable vacuum aspirator. Each sample was frozen, and pollinators subsequently were sorted and identified to the lowest genera level using various identification resources including the insect collections at the University of Connecticut as well as the online key to eastern North American bee species at www.discoverlife.org. Dr. Sam Droege of the United States Geological Survey Patuxent Wildlife Research Center assisted in the confirmation of the final identifications of pollinator species.
6. Conducted statistical analysis of pollinator visitation to determine those forage species most preferred by targeted group of native pollinators.
We used analysis of variance (ANOVA) with Tukey-Kramer adjusted means separation (PROC MIXED, SAS v 9.1) to determine differences among plant species in the number of pollinators that visit plants in early spring (March to April), late Spring (May to early June), summer (June to August), and fall (September to October) for vacuumed samples and timed observations. Simple linear regression analyses was conducted with each pair of floral characters to check for autocorrelation, and a multiple linear regression analysis was conducted on the pollinators obtained during vacuum sampling to determine bee category abundance.
7. Identified those species of pollinators that visit both crops that benefit from buzz pollination and flowering plants native to New England. Determined which pollinator species are known to provide buzz pollination.
After collecting data concerning pollinator visitation to the crops and to the common garden of native plants, we determined which pollinator species they share.
Survey and Interview Results
A survey was created using Qualtrics that explored how farmers whose crops would benefit from buzz pollination meet their pollination needs, and was sent to over 500 farmers from Connecticut, Massachusetts, and Rhode Island. We received 97 responses, from which we were able to attain agreements from six farmers to participate in our research by allowing our team to establish pollinator habitats on their farms. In addition, 33 farmers indicated they would be willing to participate in educational programs once we complete our research.
In the fall of 2017, we surveyed via email over 500 farms from Connecticut, Massachusetts, and Rhode Island and received 97 responses from farms that grow over 1500 acres of crops that would benefit from buzz pollinators. Of these responses, 57 farms consisting of over 850 acres expressed interest in establishing on-site pollinator habitats, 34 farms expressed interest in participating in educational programs concerning wild buzz pollinators, and 29 expressed interest in participating in our research trials. Our team used these surveys and the connections with farmers established by extension specialist Stearns to select four farms on which to establish on-farm wild buzz pollinator habitats.
The significant need and interests of beneficiaries was obvious from a statement one of the farmers included in his returned survey: “…I am particularly interested in building bumblebee populations. (Honeybees do little or nothing in our blueberries.)” Several farmers made similar comments concerning the return on investment they perceive they receive from honey bee use, including one of our collaborating farmers, Sandra Rose of Rose’s Berry Farm in South Glastonbury, CT, who called honey bees “lazy.” Nevertheless, a large proportion of farmers, including Rose, still depend on honey bees for pollinator services. Eighty-three, or 88.3%, of the 94 farms that responded to our surveys indicated they take some form of action to address their pollinator needs. Sixty-one, or 64.9%, of farms rent honey bee hives and 23, or 24.5%, rent or buy bumblebee quads. Of those who provided estimates for expenditures on bee rentals, the average honey bee hive rental cost $85.45 and the average bumblebee quad rental cost $76.35, resulting in these farms spending between $55 and $121 per acre annually to service crops that benefit from buzz pollination. Twenty-nine, or 30.9%, of farmers leave adjacent marginal land unmowed to benefit on-site pollinators, constituting 22 acres of possible pollinator habitat. Only 11, or 11.7%, of farms actively planted pollinator forage, whether in the form of cover crops, or native herbaceous or woody species, that cover less than eight acres in total.
Interactions with collaborating farmers also provided important feedback that helped shape how we structured pollinator habitats. A frequent concern raised by farmers was the potential weediness of some of the species, such as Monarda fistulosa and Solidago spp., included in habitats. The fact that we always intended to plant species in blocks allows for the removal of flowers once their pollen and nectar rewards have finished, thus preventing the spread of their seeds. Jamie Jones of Jones Family Farms in Shelton, CT originally was reluctant to participate when we indicated willows would be included in early blooming habitats. He called pussy willows (Salix discolor) “woody weeds.” However, he agreed to participate when we explained that we only intended to include male specimens of willows since research has found them more attractive to our predominantly pollen-harvesting buzz pollinators. We also came to realize we had to ensure that the placement of these linear plantings took into consideration the ability of large machinery to maneuver among the crops.
Finally, Craig Schatzlein, the manager of Lyman Orchards in Middlefield, CT, asked if we could create an educational program for his customers to explain what we were researching with these habitat installations. He saw this as an opportunity to show customers their efforts to implement more ecologically-sustainable practices. In addition, as a self-pick farm that attracts families, he asked if we could conduct these programs between the times that blueberry picking ends in late-July to earlyAugust and pumpkin and apple picking starts in mid-September, which usually results in a period of decreased customer traffic. Therefore, we agreed to implement a pilot program at his farm.
Early pollinator activity
Review of lists of native plants that attract pollinators in the Northeast United States not only do not include many early spring flowering plants but do not stress the importance of including varieties of such plants. In spring 2018, our team conducted preliminary research concerning
early spring pollinator activity among Salix species by placing colored – blue, white, and yellow - pan traps filled among experimental plantings and wild populations of Salix at the University of Connecticut’s Plant Science Research Farm in Mansfield-Storrs, Connecticut. The wild
populations consisted of scattered specimen of pussy willows (Salix discolor). Experimental plantings came in the form of a common garden and riparian buffer, each containing mixes of native species (S. eriocephala, S. lucida, S. nigra, S. sericea, and biofuel cultivars (S. ‘Oneida’, S.
‘Onondaga’, S. ‘S365’, S. ‘SX64’, S. ‘SX67’, S. ‘Tully Champion’) of differing bloom times. The bowls were set out for either 24 or 48 hours from March 11th to May 2nd.
We collected 384 bees, of which 296 or 77.08% were native sonicating species from the families of Andrenidae, Colletidae, and Halictidae that frequently have been documented to pollinate Vaccinium spp. of crops. Unfortunately, the pan trapping method has several known bias; in particular, they are especially attractive to halictid bees, less attractive to the genus Collettes, and infrequently attract bumblebees and honey bees, which was borne out by our own research, which resulted in only capturing one bumblebee queen and seven honey bees. These findings contributed to our altering the methodology by which we collect data on pollinator visitation and adding insect vacuuming to our methods.
Blueberry Pollinator Activity and Visitation
From May 5th to June 6th of 2018, we used both pan trap and vacuuming methods for collecting pollinator activity and visitation at four different farms that grow Vaccinium corymbosum. We categorize pollinators collected via pan traps as “pollinator activity” since we do not observe their visits to flowers. Those collected via vacuuming are categorized as “pollinator visitation” since they are vacuumed only after we witness flower visitation. Cultivars of Vaccinium corymbosum from which pollinators were collected included ‘Berkeley’, ‘Bluecrop’, ‘Blueray’, ‘Berkeley’, ‘Collins’, ‘Coville’, ‘Earliblue’, ‘Herbert’, ‘Jersey’, ‘Patriot’, and ‘Spartan’.
The composition of pollinator populations collected from pan traps differed considerably from those collected via vacuuming. We collected 761 specimens using pan traps. The percentage of genera collected via pan trapping were: Andrena spp (33.2%), Lasioglossum spp.(22.9%), Halictus spp. (11.7%), Osmia (5.3%), Collettes spp. (2.4%), and Bombus and Xylocopa (both <1%). We vacuumed 384 specimens. The percentage of genera collected via vacuuming were: Bombus queens (33.1%) Andrena spp (15.1%), Apis mellifera (11.2%), Halictus spp. (11.2%), Lasioglossum spp.(9.4%), Bombus workers (5.5%), and Xylocopa spp (4.2%).
Pollinator Forage Visitation
From June 20th to September 8th of 2018, we vacuum collected arthropod pollinators from herbaceous native flowering species planted in 2014 at UConn’s Research Farm as display plots used as part of research we conducted for New England departments of transportation that
focused on roadside revegetation using native species. Plots measured 2 sq m. We measured 1 sq m portions of these plantings and vacuumed pollinators for five minutes. We then freeze killed the specimens. Since caterpillars, butterflies, and moths do not constitute buzz pollinators, we
avoided collecting them as best we could.
By the time we finished collecting specimen in September, we concluded that five minutes was too long a period for vacuuming. For 25 summer blooming herbaceous species, we ultimately collected 5,238 specimens – a result that consumed an exorbitant amount of time to analyze and identify, and seemed a waste of the lives of these arthropods. As a result, we decided that, since we will be collecting data from over 50 species of summer pollinator forage, vacuuming times for summer blooming species would last only one minute, which is double the 30 seconds Tuell et al. (2008) used for their common garden experiment. For species that bloom prior to Vaccinium corymbosum anthesis, on the other hand, vacuum lengths will range from between five minutes to 15 minutes, depending upon the dates of anthesis and abundance of visitation
activity.
In 2019, we collected data on insect visitation from 42 species of flowering forbs from the common garden we established in 2018. The forbs species were classified into native and near-native species. Native species are those that are considered native by the Native Plant Trust’s Go Botany website, which provides information about plant species deemed native to the New England region. Near-native species are species native to North America and regions in close proximity to the New England region and frequently get included among recommendations for use for pollinator forage by conservation experts and organization. Our intention in including these near-native species was to determine if pollinators and other beneficial insects showed a preference for them that may warrant their inclusion in pollinator habitats.
As previously stated, in 2018 we collected preliminary data from display gardens as we worked to develop our methodological approach once the common garden was established. Included among our statistical analysis is data collected from 2018. There are two species of forbs for which we collected data in 2018 but did not include in the common garden - Achillea millefolium and Doellingeria umbellate – because we believed other species had attributes better suited for pollinator habitats.
Table 1. Species of forbs from which insect visitation was collected at the University of Connecticut Research Farm in Storrs CT (family: Ast: Asteraceae; Ama: Amaryllidaceae; Api: Apiaceae; Apo: Apocynaceae; Cam: Campanulaceae; Fab: Fabaceae; Lam: Lamiaceae; Phr: Phrymaceae; Pla: Plantaginaceae; Ran: Ranunculaceae; Ver: Verbenaceae; Peak Bloom Periods: E = Early (mid-May to late June), M = Middle (July to mid-August) and
L = Late (mid-August to late September); Y= native to New England, N = near-native to New England).
Table 2. Insect species and their abbreviations identified at the common garden at the University of Connecticut Research Farm in Storrs in 2018 and 2019.
|
Buzz pollinator [based on Cardinal et al (2018)] |
Non-buzz pollinators and other insects |
|
Bombus queens: BQ |
Honeybee (Apis mellifera): AM |
|
Bombus workers: BW |
Parasitic: PA (Nomada, Cuckoo bees) |
|
Xylocopa: XY |
Flies: FL |
|
Ceratina: CE |
Wasps: WA |
|
Andrena: AN |
Lady bugs: LB |
|
Colletes: CO |
Other: OT |
|
Halictus: HA |
|
|
Lasioglossum: LA |
|
|
Megachile: ME |
|
|
Osmia: OS |
|
The following categories of insects were used for statistical analysis:
Bombus buzz pollinators (BQ, BW): bomb_buzz
Non-Bombus buzz pollinators (XY, CE, AN, CO, HA, LA, ME, OS): nobomb_buzz
Non-buzz pollinator honeybees (AM, Apis mellifera): honey_bees
“Beneficial” (insect known for consuming agricultural pests) (WA, LB): bene
Other insects (PA, FL, OT): Other
These classifications are important for the following reasons:
- 75% of buzz pollination is provided by Bombus Some forb species tend to attract predominantly Bombus species while others attract a wider selection of buzz pollinators and other insects. This biodiversity in insects attracted to these plants is called richness, which adds to the overall health of agricultural ecosystem health.
- In the context of agriculture, the term beneficial frequently is applied to groups of insects that help increase the yield in crop production. For example, insects that provide pollination services have been described as beneficial. In the context of this research, the term is applied to a group of insects, including wasps and lady bugs, which tend to parasitize and decrease agricultural pests.
Descriptive Statistics
The data collected from 2019 was far more robust than that collected in 2018 (583 observations in 2019 vs 137 in 2018) because 1) we used the 2018 season to collect preliminary data for refining our methodological approach once we collected visitation data from the common garden, 2) we collected data from 43 species of forbs in 2019 compared to 16 in 2018, and 3) we standardized the number of vacuum session to a minimum of eight sessions in 2019 while our the number of sessions per forbs species varied widely. We also changed the length of our vacuuming sessions from five minutes in 2018 to one minute in 2019. Therefore, any data per observation from 2018 was divided by five.
Table 3. Means and standard deviations for number of insects per insect type, plant, and year.
|
plant |
year |
n |
honey_bees |
bomb_buzz |
nobomb_buzz |
bene |
other |
|
AchMil |
2018 |
3 |
0 (0) |
0 (0) |
0.67 (0.83) |
0.07 (0.12) |
0.07 (0.12) |
|
AgaFoe |
2018 |
10 |
0.02 (0.06) |
5.52 (1.78) |
0.3 (0.25) |
0 (0) |
0 (0) |
|
AgaFoe |
2019 |
11 |
0 (0) |
3 (3.41) |
0.55 (0.82) |
0 (0) |
0 (0) |
|
AgaScr |
2019 |
11 |
0 (0) |
5.27 (1.79) |
0 (0) |
0 (0) |
0.09 (0.3) |
|
AgeAlt |
2019 |
10 |
0.2 (0.42) |
0.7 (0.82) |
0.1 (0.32) |
0.5 (0.97) |
1 (1.15) |
|
AllCer |
2019 |
10 |
0.2 (0.42) |
1 (1.05) |
0.3 (0.67) |
0 (0) |
0.2 (0.63) |
|
AneVir |
2019 |
15 |
0 (0) |
0 (0) |
1.13 (1.55) |
0 (0) |
0.53 (0.99) |
|
AquiCan |
2019 |
14 |
0 (0) |
0.21 (0.58) |
0.79 (1.19) |
0 (0) |
0 (0) |
|
AscTub |
2018 |
9 |
0.22 (0.19) |
3.33 (1.72) |
0.13 (0.22) |
0.11 (0.15) |
0.02 (0.07) |
|
AscTub |
2019 |
10 |
0 (0) |
0.6 (0.7) |
0.2 (0.42) |
0.5 (0.71) |
0.2 (0.42) |
|
CheGla |
2019 |
15 |
0 (0) |
3.73 (1.91) |
0 (0) |
0 (0) |
0 (0) |
|
CirDis |
2019 |
12 |
0 (0) |
5.5 (1.68) |
0.08 (0.29) |
0.08 (0.29) |
0 (0) |
|
CorLan |
2019 |
13 |
1.77 (0.93) |
2.23 (1.69) |
0.31 (0.63) |
0 (0) |
0.15 (0.38) |
|
DesCan |
2019 |
13 |
0.31 (0.63) |
0.62 (1.04) |
0.23 (0.6) |
0 (0) |
2.38 (2.79) |
|
DoeUmb |
2018 |
11 |
0.05 (0.09) |
2.6 (1.62) |
0.56 (0.47) |
0.78 (0.54) |
0.22 (0.21) |
|
EchPur |
2019 |
15 |
0 (0) |
1.2 (1.37) |
0.2 (0.56) |
0 (0) |
0.4 (0.91) |
|
EryYuc |
2018 |
13 |
0.11 (0.13) |
4.02 (2.49) |
0.03 (0.08) |
0.72 (0.47) |
0.25 (0.39) |
|
EryYuc |
2019 |
11 |
0.18 (0.4) |
3.45 (2.94) |
0.36 (0.67) |
0.64 (0.81) |
1.18 (1.33) |
|
EupPer |
2018 |
7 |
0.09 (0.16) |
1.37 (1.25) |
0.23 (0.52) |
1.69 (0.86) |
0.4 (0.28) |
|
EupPer |
2019 |
13 |
0 (0) |
1.54 (1.76) |
0.38 (1.12) |
2.38 (2.63) |
0.54 (0.97) |
|
EutGra |
2019 |
8 |
0.12 (0.35) |
1.25 (0.89) |
0.12 (0.35) |
0.75 (1.04) |
0.5 (0.53) |
|
HelHel |
2018 |
8 |
0 (0) |
1.2 (0.99) |
0.12 (0.18) |
0 (0) |
0.02 (0.07) |
|
HelHel |
2019 |
12 |
0 (0) |
1.17 (1.47) |
1 (1.35) |
0 (0) |
0.5 (1) |
|
LobCar |
2019 |
12 |
0 (0) |
0 (0) |
0 (0) |
0 (0) |
0 (0) |
|
LobSip |
2018 |
9 |
0 (0) |
4.67 (2.1) |
0 (0) |
0 (0) |
0.09 (0.15) |
|
LobSip |
2019 |
16 |
0 (0) |
2.81 (2.56) |
0 (0) |
0 (0) |
0 (0) |
|
MimRin |
2019 |
9 |
0 (0) |
0.56 (0.73) |
0 (0) |
0 (0) |
0 (0) |
|
MonFis |
2018 |
10 |
0.02 (0.06) |
5.64 (3.54) |
0.06 (0.13) |
0.12 (0.25) |
0.02 (0.06) |
|
MonFis |
2019 |
18 |
0.06 (0.24) |
8.33 (5.13) |
0.06 (0.24) |
0.11 (0.47) |
0.11 (0.32) |
|
PenDig |
2018 |
3 |
0 (0) |
7.6 (1.06) |
2.8 (0.6) |
0 (0) |
0 (0) |
|
PenDig |
2019 |
16 |
0 (0) |
3.31 (4) |
0.56 (0.73) |
0 (0) |
0 (0) |
|
PenHir |
2019 |
16 |
0 (0) |
1.38 (2.09) |
0.75 (0.93) |
0 (0) |
0 (0) |
|
PenPal |
2019 |
15 |
0 (0) |
1.47 (2.83) |
0.6 (0.74) |
0.13 (0.35) |
0.67 (1.29) |
|
PycMut |
2018 |
13 |
1.06 (2.35) |
14.11 (7.27) |
0.18 (0.29) |
0.22 (0.28) |
0.02 (0.06) |
|
PycMut |
2019 |
16 |
5.25 (4.19) |
0.94 (2.49) |
1.06 (1.12) |
0.25 (0.58) |
0.5 (0.73) |
|
PycTen |
2018 |
11 |
0.29 (0.36) |
6.75 (3.83) |
0.05 (0.09) |
0.09 (0.19) |
0.11 (0.26) |
|
PycTen |
2019 |
14 |
6.57 (3.61) |
3.43 (3.06) |
0.79 (0.89) |
0.29 (0.47) |
0.93 (1.14) |
|
PycVer |
2019 |
11 |
1.55 (1.04) |
2.18 (1.99) |
0 (0) |
0.36 (0.67) |
0.45 (0.69) |
|
PycVir |
2019 |
14 |
3.71 (2.84) |
2.43 (2.31) |
0 (0) |
1.29 (1.82) |
1.79 (3.21) |
|
RudHir |
2018 |
6 |
0 (0) |
0.17 (0.08) |
0.23 (0.32) |
0 (0) |
0 (0) |
|
RudHir |
2019 |
14 |
0.21 (0.58) |
0.14 (0.36) |
0.21 (0.43) |
0 (0) |
0 (0) |
|
RudLac |
2019 |
14 |
0.21 (0.58) |
0.14 (0.36) |
0.14 (0.53) |
0 (0) |
0.43 (0.94) |
|
SenHeb |
2019 |
16 |
0.19 (0.54) |
3.12 (4.03) |
0.38 (0.81) |
0.19 (0.54) |
0.06 (0.25) |
|
SolCae |
2019 |
14 |
0.14 (0.36) |
0.71 (1.73) |
0 (0) |
0.07 (0.27) |
0 (0) |
|
SolJun |
2018 |
10 |
1.04 (1.53) |
4.18 (2.36) |
0.04 (0.08) |
0.74 (0.87) |
0.04 (0.08) |
|
SolJun |
2019 |
17 |
0.65 (1.41) |
1.47 (2.43) |
0 (0) |
0.71 (1.36) |
0 (0) |
|
SolRug |
2019 |
10 |
1 (1.41) |
1.8 (2.25) |
0.3 (0.67) |
0.7 (1.06) |
0.5 (1.08) |
|
SolSem |
2019 |
13 |
2 (1.29) |
1.08 (1.66) |
0.38 (0.87) |
0.15 (0.38) |
0.69 (0.85) |
|
SolSpe |
2019 |
12 |
2.92 (2.81) |
17.33 (8.79) |
0.5 (0.8) |
0.17 (0.39) |
0 (0) |
|
SymCord |
2019 |
12 |
0.5 (1.17) |
7.25 (6.2) |
0.5 (0.8) |
0.33 (0.49) |
0.25 (0.45) |
|
SymLae |
2019 |
14 |
1.29 (2.02) |
4 (3.46) |
0.14 (0.36) |
0 (0) |
0.5 (0.65) |
|
SymLat |
2019 |
12 |
0 (0) |
10.25 (5.99) |
1.17 (1.8) |
2.08 (2.15) |
0.67 (0.78) |
|
SymNov |
2019 |
24 |
3.54 (3.68) |
2.25 (3.6) |
0.29 (0.55) |
0.08 (0.28) |
0 (0) |
|
VerHas |
2018 |
8 |
0 (0) |
2.2 (1.04) |
1.27 (0.75) |
0.2 (0.19) |
0 (0) |
|
VerHas |
2019 |
14 |
0 (0) |
0.5 (0.94) |
0.71 (0.99) |
0.21 (0.43) |
0 (0) |
|
VerNov |
2018 |
6 |
0.13 (0.24) |
1.13 (0.53) |
0.27 (0.24) |
0 (0) |
0 (0) |
|
VerNov |
2019 |
14 |
0 (0) |
0.5 (1.4) |
0 (0) |
0.07 (0.27) |
0.07 (0.27) |
|
VerVir |
2019 |
11 |
0 (0) |
0.91 (1.14) |
0.45 (0.93) |
0.18 (0.4) |
0 (0) |
|
ZizAur |
2019 |
22 |
0 (0) |
0.73 (1.24) |
1.59 (2.28) |
0.27 (0.63) |
0 (0) |
Table 4. Means and standard deviations for number of insects per insect type and year.
year n all honey_bees bomb_buzz nobomb_buzz bene other____
2018 137 5.61 (5.03) 0.24 (0.9) 4.61 (4.67) 0.3 (0.58) 0.33 (0.57) 0.09 (0.2)
2019 583 4.52 (5.26) 0.83 (2.09) 2.36 (3.92) 0.59 (1.85) 0.27 (0.86) 0.33 (0.98)
all 720 4.73 (5.23) 0.72 (1.93) 2.79 (4.16) 0.54 (1.69) 0.28 (0.81) 0.29 (0.89)
Table 5. Means and standard deviations for insects per insect type, bloom period and year.
year bloom n all honey_bees bomb_buzz nobomb_buzz bene other____
2018 E 12 5.47 (3.42) 0.17 (0.19) 4.4 (2.47) 0.8 (1.25) 0.08 (0.13) 0.02 (0.06)
2019 E 111 2.86 (2.93) 0.21 (0.65) 1.31 (2.36) 0.87 (1.4) 0.07 (0.32) 0.18 (0.65)
2018 L 55 4.85 (2.05) 0.21 (0.74) 3.54 (2.23) 0.39 (0.58) 0.53 (0.76) 0.12 (0.2)
2019 L 197 5.96 (6.46) 0.94 (2.05) 3.6 (5.18) 0.79 (2.84) 0.28 (0.84) 0.24 (0.59)
2018 M 70 6.22 (6.62) 0.28 (1.07) 5.48 (6.03) 0.14 (0.27) 0.21 (0.36) 0.07 (0.22)
2019 M 275 4.16 (4.77) 1 (2.42) 1.89 (3.06) 0.34 (0.76) 0.35 (1) 0.46 (1.26)
all E 123 3.12 (3.06) 0.2 (0.62) 1.61 (2.54) 0.87 (1.38) 0.07 (0.31) 0.16 (0.62)
all L 252 5.72 (5.81) 0.78 (1.87) 3.59 (4.69) 0.7 (2.53) 0.34 (0.83) 0.22 (0.53)
all M 345 4.58 (5.25) 0.85 (2.23) 2.62 (4.11) 0.3 (0.69) 0.32 (0.91) 0.38 (1.14)
Graph 1. Distribution of insect type (honeybees, buzz pollinators, and non-pollinators) per plant collected in 2018.
Graph 2. The number of honeybees, buzz pollinators, and non-pollinators per plant collected in 2019.
Graph 2. The number of honeybees, buzz pollinators, and non-pollinators per plant collected in 2019.
Graph 3. Mean number of buzz pollinators per bloom period, plant species, and year in 2018.
Graph 3. Mean number of buzz pollinators per bloom period, plant species, and year in 2018.
Graph 4. Mean number of buzz pollinators per bloom period, plant species, and year in 2019.
Graph 4. Mean number of buzz pollinators per bloom period, plant species, and year in 2019.
Graph 5 presents the number of insects per insect type and plant over time. This chart illustrates that more insects were attracted to the native plant species compared to the non-native/near-native species. In addition, for the native plants, there appear to be more pollinators between the end of July/beginning of August in both 2018 and 2019, and in the end of September/beginning of August in 2019.
Graph 5. Number of insects per insect type, native vs. non-native/near-native species vs. time.
Graph 5. Number of insects per insect type, native vs. non-native:near-native species vs. time.
Conclusions:
The research focused mainly on maximizing the health and size of on-site native pollinator populations for crops requiring sonication pollination. Since it is important to have forage resources for pollinators throughout the growing season, we divided the growing season into three bloom periods: Early (E), Middle (M), and Late (L). We want to provide farmers with the top recommendations within these windows so that they can get maximum benefit for the money they invest in establishing buzz pollinator habitats. We should make clear from the beginning that our recommendations come with several caveats:
- The scope of our research focuses on data collected from the common garden we established at the University of Connecticut Research Farm.
- Our data derives mainly from one year of vacuum sessions in 2019 from the common garden we established in 2018 as well as supplemental vacuum sessions from our display gardens. We supplemented this data with analysis of the preliminary data we collected in 2018 from display garden plots, even though we used the preliminary data primarily to refine our methodology for data collection the following year. We believe it was important to include analysis of the 2018 data because we observed significant changes in the population composition of insect visitation to the same display garden plots, which influenced our final recommendations.
- We found a significant difference between the number of pollinators per plot we collected from our display gardens compared to the number we collected from those of our common garden. We assume several conditions contributed to these differences:
- The display gardens were established in 2014 and 2015. Not only did they have more time to establish, which led to a higher density of floral coverage, but we suspect that pollinators had a chance and incentive to establish nests near these plantings, thus increasing their neighboring populations.
- These display gardens included plots of warm season native bunch grasses, under which wild native ground-nesting bee species, such as Bombus, frequently nest.
- The display gardens did not have space between each species planting and the original plots were 2 m2, from which we collected from 1 m2 This greater concentration of floral resources allowed pollinators to forage from larger areas of blooms, thus allowing them to remain in one area exploiting greater floral rewards, thus expending less time and energy relearning how to access these rewards among differing floral structures.
- As stated in our introduction, we believe research into pollinator health needs to have greater focus on the distinguishing characteristics among classes of pollinators and the benefits each bring to particular agricultural crops and, consequently, recommendations for effectively maximizing the health of these various classes of pollinators depend on serving these differences in characteristics.
Early Season Blooming Species
Recommendations:
- Penstemon digitalis
- Zizia aurea
- Coreopsis lanceolata
Penstemon digitalis in 2019, by far, attracted a greater mean of Bombus pollinators (3.31) among Early blooming species compared to the mean of the other six Early blooming species (1.31). The other two Penstemon species also exceed the average mean: Penstemon pallidus (1.47) and Penstemon hirsutus (1.38). However, the mean for Coreopsis lanceolate (2.23) exceeds their means. In addition, for both Penstemon digitalis and Coreopsis lanceolate, a greater number of plants persisted from 2018 to 2019 and each plot for both species displayed a higher density of blooms per unit of area, two attributes of workhorse species. It should be noted that we classify Coreopsis lanceolate as a near-native species since it is native to areas west of New England but is often recommended by conservation organizations for pollinator forage. Our research supports such recommendations.
Zizia aurea attracted greater means of non-Bombus pollinators (1.59) and beneficial insects (0.27) among Early blooming species compared to the means of all of the other six Early blooming species (0.87) and (0.07), respectively. We believe that the biodiverse richness Zizia aurea brings to pollinator habitats merits its inclusion. In addition, it displays such workhorse attributes as persistence from year-to-year, high number of bloom per unit of area, and long bloom period.
Mid-Season Blooming Species
Recommendations:
- Monarda fistulosa
- Pycnanthemum tenufolium and Pycnanthemum muticum
- Agastache scrophularifolia
- Eupatorium perfoliatum
Monarda fistulosa in both 2018 and 2019 attracted greater means of Bombus pollinators (5.64 for 2018; 8.33 for 2019) among Mid-Season blooming species compared to the mean of the other eight 2018 Mid-Season blooming species (5.48) and twenty 2018 Mid-Season blooming species (1.89). In addition, it displays such workhorse attributes as persistence from year-to-year and high number of bloom per unit of area. It should be noted that Monarda fistulosa can spread aggressively, thus possibly causing weed problems among the main crop.
Our data for Pycnanthemum tenufolium and Pycnanthemum muticum from 2018 compared to 2019 showed enough anomalies that made it difficult for us to recommend one over the other. Our data concerning Bombus visitation from 2018 would have recommended Pycnanthemum muticum (14.11) over Pycnanthemum tenufolium (6.75) because of how they compared to the eight 2018 Mid-Season blooming species (5.48). However, something unusual happened in 2019. For both years we collected from both our display garden plots as well as those from our common garden. From year-to-year, the proportion of honeybees compared to Bombus visitations in the display garden got reversed for both species. Among 2018 Mid-Season blooming species, the honeybee mean was 0.28 compared to 1.06 for Pycnanthemum muticum and 0.26 for Pycnanthemum tenufolium. Among 2019 Mid-Season blooming species, the honeybee mean was 1.00 compared to 5.25 for Pycnanthemum muticum and 6.57 for Pycnanthemum tenufolium. Then our data concerning Bombus visitation from 2019 for Pycnanthemum muticum and Pycnanthemum tenufolium decreased to 0.94 and 3.43, respectively, compared to 1.89 for all 2019 Mid-Season blooming species. We’re not quite certain what factors contributed to such a drastic change in pollinator population composition at that point in the season for those particular patches of forbs. Extensive research exists on the competition between honeybees, which are not native to North America, and wild North American native bee populations. Initially we thought such competition was overblown following the 2018 season. However, our 2019 data raised questions anew on the subject. We first suspected that the weather at the time that Bombus species grow their broods in early spring might have been a contributing factor, but realize we would have to compare year-over-year data to start to make such determinations. Nevertheless, we believe both species would be beneficial and attractive to buzz pollinators, despite the conflicting data, In addition, Pycnanthemum species tend to have especially long bloom periods, with that of Pycnanthemum muticum extending to almost two months long. They display other workhorse attributes such as persistence from year-to-year and high number of bloom per unit of area.
Among the two Agastache species for which we collected data, we recommend Agastache scrophularifolia over Agastache foeniculum. First, Agastache scrophularifolia is native to the New England region. Secondly, although Agastache foeniculum displayed significant attractiveness when we collected data in 2018 compared to the mean for Mid-Season blooming species, we believe that was influenced by their being situated among other potted forbs we had grown that year, thus creating a desirable foraging space. When compared within the common garden space in 2019, Agastache scrophularifolia had a significantly higher mean for Bombus visitation (5.27) to that of Agastache foeniculum (3.00).
Finally, we include Eupatorium perfoliatum among our recommendations because it had the highest mean for attracting beneficial insects among Mid-Season Species (2.38 vs 0.21). They also display such workhorse attributes as persistence from year-to-year and a high number of bloom per unit of area.
Late Season Blooming Species
Recommendations:
- Solidago speciosa
- Symphyotrichum cordifolium and Symphyotrichum lateriflorum
- Cirsium discolor
We planted several species of Solidago because we wanted not only to collect pollinator visitation data but examine the aggressiveness with which each species would spread. Some Solidago species are known to spread quite aggressively, frequently outcompeting other forbs and creating near monoculture stands as evidenced by their tendency to dominate roadsides in the autumn. We planted common garden plots of Solidago nemoralis and Solidago flexicaulis, two species of shorter statue from which we hoped to collect promising data. Unfortunately, they did not establish well and did not provide sizable enough plots from which to collect data. Although we collected promising preliminary data from Solidago juncea in 2018, we saw a decline in visitation in 2019 of buzz pollinators. Since Solidago juncea blooms much earlier than most Solidago – thus earning its common name early goldenrod – we had hoped to include it among our Mid-Season Blooming group of forbs. However, we probably would have excluded it because of its tendency to spread aggressively. Ultimately, we found that Solidago speciosa to be the most attractive forbs species among all the species from which we collected data. It had a mean of 17.33 for Bombus species visitation compared to the mean of 3.6 for all fifteen species from which we collected data in 2019. In addition, it displays such workhorse attributes as persistence from year-to-year and a high number of bloom per unit of area. Best of all, we found that Solidago speciosa did not exhibit a tendency to spread aggressively, as evidenced by the 2m2 plot we planted in 2014, which remained mostly confined within in its original parameters.
We always expected we would include an aster species as part of our recommendations but we were surprised by those we ultimately recommended. With the common name New England aster, we expected Symphyotrichum novae-angliae to be among the more attractive aster species. While its mean for honeybees was the highest among Late Season blooming species (3.54 vs 0.94 for all Late-Season species), it had the lowest mean for Bombus visitation (2.25) among the four Symphyotrichum species from which we collected and performed below the mean for Bombus visitation among all Late-Season species (3.6). In addition, Symphyotrichum novae-angliae displayed one of the highest tendencies to spread aggressively, ultimately overtaking large portions of one of our display gardens along with Solidago juncea. One of our two final aster recommendations, Symphyotrichum cordifolium, comes with a few caveats. With a mean of 7.25 for Bombus species visitation, Symphyotrichum cordifolium attracts more than double the mean for Late Season species (3.6). However, we noticed it did not compete well with Symphyotrichum novae-angliae for persistence in our display garden. We believe this is possibly related to the fact that it thrives better with a bit of shade during the hotter parts of the day, which might be expected for a species with the common name blue wood aster. We believe this trait may be beneficial for those farmers for which part of their pollinator habitat borders a woodland edge. A bigger surprise among our aster recommendations was the attractiveness we observed for Symphyotrichum lateriflorum, a species we had not included among our common garden plantings but had established volunteer plots within the parameters of our common garden. Not only did it attract the second highest mean of Bombus species among Late Season species (10.25 vs 3.6), it had the highest mean for non-Bombus buzz pollinators among Late Season species (1.17 v. 0.79). In addition, it proved especially attractive to wasps, resulting in the highest mean for beneficial insects for Late Season species (2.08 vs 0.28).
Finally, we recommend Cirsium discolor because of its high mean for Bombus species visitation (5.5 vs 3.6 for all Late Season species). We noticed that Bombus species would linger on Cirsium discolor blooms longer than most every other forbs species we observed. Often mistaken for the non-native, invasive Cirsium vulgare, Cirsium discolor does not spread as aggressively and can co-exist among other native forbs species.
Education & Outreach Activities and Participation Summary
Participation Summary:
We are currently in discussion with the organizers of CT NOFA's 38th Annual Winter Conference to be held pm March 7th, 2020 in Middletown CT at Wesleyan University to present our findings. We will include distribution of our two fact sheets as part of outreach to farmers.
FACT SHEETS
Fact-Sheet-Constructing-Bee-Habitats-for-Crops-Benefiting-Buzz-Pollinators
Fact Sheet - Selection of Preferred Forage for Buzz Pollinators
Project Outcomes
We have created two Fact Sheets that help explain to farmers who grow crops that benefit from buzz pollination that their crop yields can increase if they serve the needs of those species of pollinators that more effectively pollinate their crops. As we learned from our surveys, these farmers are spending significant amounts of resources on honeybees for which many of them admitted they are not receiving a great enough return on their investments. However, most of them did not understand why. We hope that these fact sheets will help inform their choices when making decisions concerning how to meet their pollination needs, especially when it comes to buzz pollinator forage resources and habitat structure.
We discovered that competition among honeybees and bumblebees may be more significant than we originally believed. When we collected data concerning pollinator forage preference in 2018, bumblebees constituted the vast majority of pollinators collected. However, from the same plots of forbs, we found the proportion of honeybees to bumblebees within samples collected reversed themselves in terms of proportions.
In addition, while working with blueberry farmers, we realized that their needs and habitats had to considered as much, if not more, when constructing habitats. For example, while having foraging resources of native flowers benefit wild bee populations, these native species can create weed pressure on the produce crops. It was important to provide farmers guidance for effectively and economically managing the seeds produced by the species constituting these pollinator habitats.
We found that our proposed methodologies and promised deliverables were far more ambitious than could reasonably be conducted and produced under the funding constraints of this grant. While we applied for four grants to conduct our research as originally envisioned, we did not secure these grants because, while reviewers considered our approach sound, we could not find a sufficient number of blueberry farms large enough to allow us to replicate our treatments effectively and thus produce statistically significant results. While we grew over 12,000 plants and had started establishing 12 habitats on four blueberry farms, we could not afford the labor or travel expense to complete the habitats and to collect the pollinator visitation data at the habitats.
We also found that the composition of the population of pollinator species that visited our display gardens from one year to the next were so significantly different that buzz pollinator forage preferences deserve further research conducted over a greater number of years. In addition, data concerning nesting resources and weather during the early spring breeding period for Bombus species needs to be included as part of research to determine their impacts on pollinator population composition and size.
Finally, we believe that these farmers would benefit from tools that would calculate the costs of various methods for meeting their need for pollinator services and their ability to increase crop yields. Such a tools would help determine if these farmers could reasonable expect a return on their investment in pollinator habitats.
Works Cited
Cardinal S, Buchmann SL, Russell AL. 2018. The evolution of floral sonication, a pollen foraging behavior used by bees (Anthophila). Evolution 72: 590-600.
Tuell, JK, Ascher, JS, Isaacs, R. Wild Bees (Hymenoptera: Apoidea: Anthophila) of the Michigan Highbush Blueberry Agroecosystem. Annals of the Entomological Society of America, Volume 102, Number 2, March 2009: 275-287.