Skip to main content

Data from: Pollinators and plant nurseries: how irrigation and pesticide treatment of native ornamental plants impact solitary bees

Cite this dataset

Cecala, Jacob; Wilson Rankin, Erin (2021). Data from: Pollinators and plant nurseries: how irrigation and pesticide treatment of native ornamental plants impact solitary bees [Dataset]. Dryad.


A key conservation goal in agroecosystems is to understand how management practices may affect beneficial species, such as pollinators. Currently, broad gaps exist in our knowledge as to how horticultural management practices, such as irrigation level, might influence bee reproduction, particularly for solitary bees. Despite the extensive use of ornamental plants by bees, especially little is known about how irrigation level may interact with insecticides, like water-soluble neonicotinoids, to influence floral rewards and bee reproduction. We designed a two-factor field cage experiment in which we reared Megachile rotundata (Fabricius) (Hymenoptera: Megachilidae) on containerized ornamental plants grown under two different irrigation levels and imidacloprid treatments (30% label rate dosage of a nursery formulation or an untreated control). Lower irrigation was associated with modest decreases in nectar volume and floral abundance in untreated plants, whereas irrigation did not affect plants treated with imidacloprid. Furthermore, higher irrigation decreased the amount of imidacloprid entering nectar. Imidacloprid application strongly reduced bee foraging activity and reproduction, and higher irrigation did not offset any negative effects on bees. Our study emphasizes the impact of a nursery neonicotinoid formulation on solitary bee foraging and reproduction, while highlighting interactions between irrigation level and neonicotinoid application in containerized plants themselves.


We examined the effects of imidacloprid application and irrigation level on floral nectar and ALCB reproduction using a fully crossed randomized block design. From March 2018 to July 2020, we maintained 20 field cages (each 5.8 m3) in a 0.30-ha plot at University of California Riverside Agricultural Operations (33.965 °N, 117.341 °W). We used four cages for the nectar experiment and 16 for the ALCB reproduction experiment. Each cage served as a replicate mesocosm simulating conditions at a containerized nursery.


Nectar experiment

      We began growing lacy phacelia (Boraginaceae: Phacelia tanacetifolia) from seed in UC Soil Mix III ( in 2-L pots in a greenhouse on 29 October 2019. We focused on containerized plants, as soil dynamics for plants in containers likely differ from those experienced by plants growing in the ground. We selected phacelia due to its attractiveness to bees, including ALCBs [32], and its abundant floret production [33]. We assigned 180 plants to one of six treatments resulting from the crossing of irrigation level and imidacloprid treatment (table 1), organized into two experimental blocks of two cages each (figure S1a), resulting in 45 plants per cage.

Plants were moved into cages on 14 January 2020, and we inserted an individual high- or low-flow irrigation spike (Primerus Products, Encinitas, CA) into each pot. These spikes are widely used in nurseries for container irrigation, and we selected spikes representing the lowest and highest flow rates for 2-L pots. High-flow spikes emitted 2.6 times more water as low-flow spikes (figure S2), resulting in 23% higher average midday volumetric water content (VWC; figure S3a). All pots were automatically irrigated simultaneously over the soil surface once per day at 0600 h for 60 s, increased to 120 s whenever any plants visibly wilted due to heat. The ground inside each cage was lined with fabric barrier to prevent plants from rooting outside pots.

For imidacloprid treatments, we applied Marathon® 1% Granular (OHP, Bluffton, SC), a commercial nursery formulation, to pots on 28 January 2020. Formulated for use in greenhouses and nurseries, Marathon® consists of 1% imidacloprid and 99% inert ingredients by mass. Granular formulations allow the active ingredient to leach more slowly from potting soil than liquid drenches [34]. The label rate dosage for a 2-L pot equates to 1.4–2.0 g formulation. We used ~30% the label rate as our “high” dosage to be conservative, as near total ALCB mortality occurred when separate plants were treated at label rate (unpublished data). In this experiment, we also included a “low” dosage treatment of ~3% the label rate (table 1).

Phacelia bloom began in early May 2020 and lasted six weeks. On each of six separate days during the bloom period, we measured floral nectar volume between 1030 and 1330 h (daily high temperatures between 22.2 to 36.1 °C) in randomly selected flowers in each treatment using microcapillary tubes (Drummond Scientific, Broomall, PA) and handheld calipers (N = 13 flowers x 4 cages x 6 days = 312 flowers). We also quantified total sugar concentration of samples on three of these days using a refractometer (Eclipse, Bellingham + Stanley) (N = 7 to 8 flowers x 4 cages x 3 days = 94 flowers). On each of eight separate days during bloom, we collected nectar during the same time of day, pooled samples within treatments within cages (N = 3 pooled samples x 8 days = 24 samples (4 per treatment), and quantified imidacloprid residues in nectar via an ELISA. We used a QuantiPlate kit (EnviroLogix, Portland, ME) and microplate spectrophotometer (Thermo Fisher, Waltham, MA), which provide similar quality data to HPLC/MS-MS at less cost [16]. Imidacloprid metabolites, also toxic to bees [29], cross-react in the assay. Thus, assay results reflect the total concentration of the parent compound and its metabolites. Samples were diluted 10- to 100-fold before analysis as needed to complement the kit’s quantification range.


ALCB reproduction experiment

In the other 16 field cages, we placed the California-native ornamental plants Erigeron glaucus, Sphaeralcea ambigua, and Baileya multiradiata in 2-L pots (not previously treated with insecticides), purchased from a local native plant nursery (Valley Center, CA) in 2019 and 2020. We selected these species based on their popularity at nurseries, drought tolerance to ensure bloom in low irrigation conditions, and from surveys of wild bee visitation at nurseries (Cecala and Wilson Rankin, in review). We also included pots of lacy phacelia and alfalfa (Medicago sativa). The ratio of plant species in each cage varied slightly between years due to availability (table S1). We divided field cages into four experimental blocks, with one cage assigned to each of four treatments (table 1, figure S1b). The low dosage (~3% label rate) imidacloprid treatment was not included in the bee reproduction experiment.

We manipulated plant irrigation levels as in the phacelia nectar experiment. Although soil VWC varied across plant species, VWC in high irrigation pots was 44% higher than that of soil in low irrigation pots (figure S3b). In imidacloprid treatment cages, we applied Marathon® four weeks before introducing bees to allow for translocation [as in 14]. We applied Marathon® to each plant species except alfalfa, which we anticipated would serve as the principal leaf clipping source for nesting female ALCBs, to ensure bees were primarily exposed to imidacloprid through consumption of pollen and nectar, and not via leaf tissue clipping.

In each cage, we provided one nest block constructed according to USDA-ARS specifications [35] facing southeast. Each block contained 60 drilled tunnels into which we inserted paper straws (diameter 5 mm, length 12.7 cm). We purchased commercially reared ALCB pre-pupae from Canada (JWM Leafcutters, Parkside, SK, Canada) and allowed them to develop into adults in an incubator at 30.3 ± 0.1 °C, 57.2 ± 0.4% relative humidity (mean ± s.e.) and a 12-hour light-dark cycle. Emergence occurred after 21 days. In mid-June 2019 and 2020, we introduced 30 male and 20 female ALCBs inside each cage to approximate sex ratios in commercial populations [36].

Over the following six weeks, two to three times per week, we recorded floral abundance (for each plant species and the entire cage) and ALCB foraging activity in each cage using ordinal indices. For floral abundance, we assigned: ‘0’ if no flowers were present, ‘1’ if a few flowers were present, ‘2’ if flowers covered 10–50% the cage area, and ‘3’ if flowers covered > 50% the cage area. For bee foraging activity, we visually monitored the inside of each cage for 10 seconds [similar to 32] and assigned: ‘0’ if no foraging bees were visible, ‘1’ if 1–3 bees were visible, ‘2’ if 4–10 bees were visible, and ‘3’ if > 10 bees were visible. We also recorded ambient temperature during observations.

After six weeks, we collected all straws and individually labeled and weighed each. After three weeks of storage at 22 °C, straws were kept at 5 °C over the winter. After at least four months, straws were incubated again. However, no bees emerged in either year, potentially due to the lack of a fluctuating temperature regime during cold storage [37], insufficient quantity or quality of pollen provisions, or adults not entering diapause (though we did not observe any non-diapausing adults emerging) [38]. To assess reproduction, we dissected straws and inspected the contents, quantifying incomplete cells (leaf pieces not fully formed into a cell), empty cells (fully constructed, but with no contents), cells containing pollen provisions but no brood, and cells with brood (and the developmental stage).


Statistical analysis

We conducted all analyses in R (version 3.3.3) [39]. All means are reported ± s.e. In all models, we checked for collinearity using function ‘vif’ (car) [40]. To assess how treatments impacted volume, sugar concentration, and imidacloprid concentration of phacelia nectar, we constructed linear mixed models (LMMs) using function ‘lmer’ (lme4) [41]. We included as fixed effects irrigation level and imidacloprid treatment (and their interaction) and number of days elapsed since imidacloprid application. In the volume and sugar concentration models only, we included daily high temperature, known to influence nectar secretion in phacelia [33]. Cage nested within block served as random effects. Volume and imidacloprid concentration were log10(x+1)-transformed. In all models, we used function ‘emmeans’ (emmeans) [42] for post-hoc comparisons (Tukey’s HSD tests) as appropriate.

To assess treatment effects on indices of cage-level floral abundance and ALCB foraging activity, we constructed additional LMMs. We included as fixed effects irrigation level, imidacloprid treatment, and year (and all interactions), and number of days elapsed since bees were added to cages. For the bee foraging activity model, we also included ambient temperature during the observation. We again included cage nested within block as random effects. To assess how treatments influenced nest initiation by ALCBs, we constructed a generalized linear mixed model (GLMM) using function ‘glmer’ (lme4) and a logit link. In this model, we treated each straw (empty or not) as a replicate, noting whether there was any evidence of nest construction or not. We included the same fixed and random effects as the bee foraging activity model. Furthermore, to determine if treatments influenced (per cage) the number of cells (incomplete or complete) constructed, number of cells containing brood, or proportion of cells containing brood, we constructed LMMs with the aforementioned predictor variables. We square-root transformed the number of cells per cage and number of cells containing brood per cage. Finally, to determine if female ALCBs clipped plants other than alfalfa (which would further expose them to imidacloprid), we used Fisher’s exact tests. We tested if the number of cells (incomplete or complete) constructed with versus without alfalfa (never treated with imidacloprid) differed between imidacloprid or irrigation treatments.

Usage notes

"phacelia_nectar_data.xlsx" contains data from the experiment examining the effects of irrigation level and imidacloprid dosage on Phacelia tanacetifolia nectar characteristics.

"alcb_reproduction_data.xlsx" contains data from the experiment examining the effects of irrigation level and imidacloprid dosage on foraging and reproduction of alfalfa leafcutting bees.

Descriptions of tabs and variables are provided in the first tab (metadata) of each Excel file.


National Institute of Food and Agriculture, Award: 2019-67011-29512

California Association of Nurseries and Garden Centers, Award: NA