Adult monarch butterflies show high tolerance to neonicotinoid insecticides
Abstract
- Numerous studies have documented the negative effects of neonicotinoids on bees; it remains crucial to examine how neonicotinoids affect other non-target nectar-feeding insects, such as the monarch butterfly, Danaus plexippus.
- Wildflowers growing near agricultural areas can be contaminated with neonicotinoids that affect survival or cause sublethal changes to behaviours of nectar-feeding insects. Nectar residues of imidacloprid and clothianidin found in milkweeds and wildflowers adjacent to agricultural field range from 0 to 72.8 ng/mL.
- At field-relevant doses, two neonicotinoids (imidacloprid and clothianidin) were studied for their effects on adult monarch survival, reproduction, flight and behaviour. First, we fed adult monarchs artificial nectar solutions ranging from 15 to 386 ng/mL of imidacloprid and 19 to 531 ng/mL of clothianidin. Neonicotinoid ingestion slightly reduced monarch reproduction but had no significant effects on survival, weight change, or activity levels.
- Second, we fed monarchs higher clothianidin doses (909 and 4030 ng/mL), that exceed field-relevant levels by 22 and 99 times. These higher doses reduced monarch nectar consumption, survival, flight performance and reaction time in response to a drop test.
- Results show that adult monarchs tolerate field-relevant doses as high as 54 ng/mL for imidacloprid and 75 ng/mL for clothianidin, with minimal lethal or sub-lethal effects until much higher doses are supplied. We conclude that adult monarchs are more tolerant of ingested clothianidin and imidacloprid than indicated by previous research.
INTRODUCTION
Monarch butterflies (Danaus plexippus) are charismatic and iconic insects, with scientific interest in their biology stemming from their long-distance yearly migrations, chemical interactions with milkweed host plants and recent analyses of population declines and the mechanisms responsible (Brower, Fink, & Walford, 2006; Gustafsson, Agrawal, Lewenstein, & Wolf, 2015; Malcolm, 1994; Oberhauser, Nail, & Altizer, 2015; Zhan et al., 2014). Declines in winter colonies of North American monarchs have caused concern for the persistence of their migration (Brower et al., 2012; Schultz, Brown, Pelton, & Crone, 2017) and long-term population viability (U.S. Fish and Wildlife Service, 2020). While some evidence suggests these declines stem from issues related to habitat loss during summer (Flockhart, Pichancourt, Norris, & Martin, 2015; Pleasants & Oberhauser, 2013; Semmens et al., 2016; Thogmartin et al., 2017; Vidal & Rendón-Salinas, 2014), other research points to problems faced during the fall migration (Inamine, Ellner, Springer, & Agrawal, 2016; Ries, Taron, & Rendón-Salinas, 2015; Saunders et al., 2019) and insecticide contaminated food plants (Olaya-Arenas, Hauri, Scharf, & Kaplan, 2020; Olaya-Arenas, Scharf, & Kaplan, 2020; Prouty, Barriga, Davis, Krischik, & Altizer, 2021). In particular, common insecticides, including widely-used neonicotinoids, have been suggested as a potential cause of decreased monarch migration success (Stenoien et al., 2018; Tracy, Kantola, Baum, & Coulson, 2019). Recent work highlighted a correlation between increased use of neonicotinoids and declines in eastern North American monarchs (Forister et al., 2016; Stenoien et al., 2018; Thogmartin et al., 2017; Gilburn et al., 2015). Other work showed that both pesticide use and habitat loss predicted western North American monarch declines (Crone, Pelton, Brown, Thomas, & Schultz, 2019).
Neonicotinoids are the most widely used class of insecticide in the world, with prevalent use throughout North America on row crops, orchards, vegetables and ornamental plants (Bonmatin et al., 2015; Van der Sluijs et al., 2013). Neonicotinoids bind to nicotinic acetylcholine receptors in insects, causing paralysis and death (Simon-Delso et al., 2015) and are highly effective against many sucking, leaf chewing and soil insects (Gervais, Luukinen, Buhl, & Stone, 2010). The compounds persist in the environment for many months, with half-lives from 100s to 1000s of days in the absence of exposure to UV light (Mohapatra et al., 2019). Soils surrounding experimentally treated seeds contained between 2.1 and 9.6 ng/mL of clothianidin on average and talc samples taken from planters with clothianidin-exposed seeds contained up to 15.03 ng/mL of clothianidin (Krupke, Hunt, Eitzer, Andino, & Given, 2012). This can lead to neonicotinoid contamination in the pollen and nectar of agriculturally-adjacent plants used by beneficial insects (Krupke et al., 2012; Krupke & Long, 2015; Krupke, Holland, Long, & Eitzer, 2017; Rundlöf, 2015, Hladik, Vandever, & Smalling, 2016, Long & Krupke, 2016).
Monarchs can be exposed to neonicotinoids when their primary larval host plants (milkweeds) grow in field margins, as well as when adults forage on wildflowers in agricultural landscapes treated with neonicotinoids. One study showed milkweed leaves in agricultural sites had mean concentrations of 40.76 ng/g clothianidin and 0.46 ng/g imidacloprid (Halsch et al., 2020). Nectar of wildflowers in field margins has been shown to contain neonicotinoids and 97% of neonicotinoids brought back to honeybee hives was from wildflowers, not crops (Botías et al., 2015; 2016). Honey bee-stored nectar and honey were shown to have a maximum of 10.1 and 72.8 ng/g of clothianidin and imidacloprid, respectively (Sanchez-Bayo & Goka, 2014) and honey bee-collected pollen showed mean levels of 9.4 and 19.7 ng/g of clothianidin and imidacloprid, respectively (Sanchez-Bayo & Goka, 2014). There are contexts under which monarchs and other pollinators could become exposed to higher doses than field margins. Cowles and Eitzer (2017) found that milkweed treated with neonicotinoids products commonly bought by gardeners can lead to doses up to 1000 ng/mL in nectar. Plants in some retail nurseries have insecticides at high concentrations; for example, Halsch et al. (2020) found mean concentrations of cyantraniliprole at >500 ng/mg, together with other types of insecticides such as bisamides, diacylhydrazines and neonicotinoids. In another retail plants study (Halsch, Hoyle, Code, Fordyce, & Forister, 2022), clothianidin was found at up to 2.58 ng/g and imidacloprid at up to 37.69 ng/g. Nectar from directly-sprayed clover had up to 2992 ng/g and clothianidin and 6588 ng/g of imidacloprid, but concentrations were reduced by 99.4%–99.8% in blooms after mowing (Larson, Redmond, & Potter, 2015, reviewed in Zioga, Kelly, White, & Stout, 2020). If gardeners and nurseries use neonicotinoid products to deter pest insects, monarchs and other pollinators could be exposed to high doses via this route.
The negative fitness effects of neonicotinoid insecticide ingestion have been well documented in bees, but more work is needed to examine neonicotinoid effects across a broader range of non-target insects. In bees, at low concentrations (<50 ng/mL, readily found in field samples) bees suffer from reduced foraging efficiency and poor nutritional status (Azpiazu et al., 2019; Feltham, Park, & Goulson, 2014; Lämsä, Kuusela, Tuomi, Juntunen, & Watts, 2018; Morfin, Goodwin, Correa-Benitez, & Guzman-Novoa, 2019; Phelps, Strang, & Sherry, 2020; Scholer & Krischik, 2014; Stanley & Raine, 2016; 2017). There is also evidence for low-dose effects on bee social behaviour, reproduction, navigation and flight performance (Bryden, Gill, Mitton, Raine, & Jansen, 2013; Crall et al., 2018; Fischer et al., 2014; Laycock, Lenthall, Barratt, & Cresswell, 2012; Scholer & Krischik, 2014; Switzer & Combes, 2016; Tosi, Burgio, & Nieh, 2017; Whitehorn, Wallace, & Vallejo-Marin, 2017). In recent years, a number of studies tested neonicotinoid effects on the larval stage of butterflies, where ingestion occurs via insecticide in host plant material (e.g., Krishnan et al., 2020, Krishnan et al., 2021, Krishnan, Jurenka, & Bradbury, 2021, Olaya-Arenas, Hauri, et al., 2020, Knight, Flockhart, Derbyshire, Bosco, & Norris, 2021, Bargar, Hladik, & Daniels, 2020 and summarised in Prouty et al., 2021). A much smaller number of studies have tested the consumption of neonicotinoids in monarchs as adults, with mixed results. Adult monarchs fed ad libitum with a field-relevant concentration of imidacloprid (23 ng/mL), showed low survival (James, 2019). In contrast, adult monarchs fed nectar containing 15 and 30 ng/mL imidacloprid (Krischik, Rogers, Gupta, & Varshney, 2015), 140 ng/mL clothianidin, 250 ng/mL imidacloprid and 330 ng/mL thiamethoxam (Krishnan, Zhang, et al., 2021) showed no reduction in survival, despite higher concentrations than found in field nectar residues (Botías et al., 2015; 2016).
Our objectives were to examine how imidacloprid and clothianidin consumption by adult monarchs influenced their survival, behaviour, reproduction and flight performance. We first exposed monarchs to different concentrations of imidacloprid and clothianidin (15–531 ng/mL, with a control for each neonicotinoid), via artificial nectar over multiple days. After finding no effect of either compound at these doses on monarch survival and only weak effects on reproduction, we conducted a second experiment with higher doses of clothianidin (909 and 4030 ng/mL) and recorded monarch survival and flight performance. We predicted that imidacloprid and clothianidin would reduce monarch reproduction, as shown in bees (e.g., Chan, Prosser, Rodríguez-Gil, & Raine, 2019; Chan & Raine, 2021; Dolezal, 2022; Laycock et al., 2012; Stuligross & Williams, 2021), since mating and egg-laying require energetically expensive complex behaviours (e.g., Brower, Oberhauser, Boppré, Brower, & Vane-Wright, 2007; Oberhauser, 1989). Also, we predicted that monarch flight speed and distance would be reduced following the ingestion of clothianidin and imidacloprid.
MATERIALS AND METHODS
Neonicotinoid use and residue analysis
Clothianidin and imidacloprid were obtained from Sigma-Aldrich. Each neonicotinoid (1 mg) was dissolved separately into 0.5 L of distilled water to achieve 2 ppm (mg/L) stock solutions for experiment 1; 5 mg of clothianidin was dissolved into 0.5 L to achieve a 10 ppm (mg/L) stock solution for experiment 2. Stock solutions were diluted with a honey-water solution (distilled water and honey to maintain a 20% (v/v) honey concentration) to intended doses of 25, 50, 100 and 500 ppb (ng/mL) imidacloprid and clothianidin for experiment 1; and 1000 and 5000 ppb (ng/mL) clothianidin for experiment 2. Stock solutions were held at 4°C for up to 7 days prior to each experiment and dosed honey solutions were prepared fresh every 3 days during feeding trials.
Aliquots of artificial nectar (6–10 mL) were stored in scintillation vials at −20°C and mailed to the USDA, Agricultural Marketing Service, Science Laboratory Approval and Testing Division, Gastonia, NC. Samples were extracted using a refined methodology for the determination of neonicotinoids using the official pesticide extraction method (AOAC OMA 2007.01) and analysed by liquid chromatography (LC) coupled with tandem mass spectrometry detection (LC/MS/MS). Quantification was performed using external calibration standards prepared from certified reference material. The instrumentation used was an Agilent LC/MS/MS (1200 pumping systems and models 6420 or 6430 detection units). Each 3 g subsample of a homogenised sample is fortified with one or more process control standards (PCS). The analytes of interest and PCS(s) are extracted from the samples by high-speed grinding in an acidified acetonitrile and water mixture followed by a “clean-up” to remove some matrix components and filtration to remove particulates. Enhance matrix reduction (Agilent EMR) material is utilised to remove lipids from applicable matrices. Separate aliquots of extract are analysed for pesticide residue by gas chromatography (GC) and LC techniques utilising mass selective detection systems. Limits of detection were 3 ng/mL for clothianidin and 2 ng/mL for imidacloprid.
Wild-Caught bumble bee bioassay to confirm spiked solution dose
To confirm the toxicity of the neonicotinoid treatments used here, wild-caught Bombus impatiens were exposed within 1 h of capture to clothianidin: 0, 19, 36, 75, 531 ng/mL or imidacloprid: 0, 15, 28, 54, 386 ng/mL. Bees were captured while foraging at the UGA campus in Athens, GA USA. On average, the bumblebees used for this bioassay were of similar weight to monarchs (0.3–0.8 g) at the time of exposure. We used the same stock solutions as for Experiment 1 to prepare a 20% spiked honey water solution for each neonicotinoid treatment. Bees were placed into clear plastic containers (15 × 10 × 10cm) with mesh screen lids, to which we added sponges soaked in the honey water solution to petri dish bottoms (n = 4 bees/container, 1 container/treatment, 40 bees in total). Containers were checked every 40–60 min over a 4 h period to record the activity of each bee as follows: flying, walking/crawling, standing, twitching and lying on side, or dead (Table S1). Bees were frozen at −20°C after observations were concluded.
Monarch sources
For Experiment 1, we used lab-reared monarchs from 5 outcrossed genetic lineages that were the descendants of ~100 wild-caught fall migrants from Athens, GA and St. Marks, FL, USA collected in Oct 2018. To obtain eggs, adult monarchs were housed in 0.6 m3 mesh cages for mating and fed ad libitum with a 20% (v/v) honey-water solution. Eggs were laid on stalks of milkweed (Asclepias incarnata). Plants were raised from seed (Prairie Moon Nursery, MN) and planted into 12.5 cm diameter pots in February 2019 in a climate-controlled greenhouse (range 15–35°C). Plants were cut back several times prior to the study and received bi-monthly Osmocote fertiliser. Larvae remained on natal stalks until second instar. Monarch larvae were reared individually in 0.6 L plastic containers with mesh screen lids on milkweed cuttings at 25–28°C and exposed to natural light in a windowed room. Containers were arranged on shelves and monitored twice daily for death, pupation and eclosion. After 24-h post-eclosion, we recorded monarch sex and weighed monarchs to the nearest 0.001 g and checked for infection by the protozoan, Ophryocystis elektroscirrha (Altizer et al. 2000), which can negatively affect flight and survival.
For Experiment 2, wild adult monarchs were collected in Athens, Georgia in October and November 2019. Monarchs that tested positive for O. elektroscirrha were excluded from further study. Monarchs were kept in an incubator to simulate day length and temperatures at the Mexico overwintering sites to maintain reproductive diapause and simulate overwintering. Adults were fed a 20% honey water solution every 11–12 days for 4 months. Of the 100 wild-caught monarchs, 40 individuals (20 males and 20 females) were randomly selected for neonicotinoid exposure and flight trials. We used an additional 12 lab-reared monarchs (4 per treatment group; six female, six male), raised under conditions described in Experiment 1, to determine whether wild-caught and lab-reared monarchs respond similarly to neonicotinoids.
Experiment 1: effects of low dose neonicotinoids on movement and reproduction of lab-reared monarchs
To test the effect of neonicotinoid consumption in nectar on adult monarchs, we fed adults every other day for 10 days, using 20% honey water solutions (v/v) mixed with clothianidin (0, 19, 36, 75, 531 ng/mL) or imidacloprid (0, 15, 28, 54, 386 ng/mL), n = 36 monarchs/treatment. Feeding was accomplished by restraining monarchs with steel nuts on plexiglass feeding trays (Figure S1) for 10 min per treatment. We weighed monarchs to the nearest 0.001 g before and after each feeding to determine the amount ingested. We used separate trays for each dose-by-neonicotinoid type treatment to minimise cross-contamination and sanitised trays by exposure to artificial UV light for 1hr at the end of each day followed by soaking overnight in 20% bleach solution. Controls (honey-water only) were fed in an adjacent room on separate trays to limit the potential for neonicotinoid exposure. All adults were stored in individual glassine envelopes at 23°C (14 h daylength) in the same incubator between feedings.
Following the fifth and final feeding, we randomly assigned butterflies from each treatment, including controls, to 0.6 m3 indoor mesh cages (Figure S2). We used a total of 16 cages with approximately 20 butterflies per cage. Butterflies were numbered using ultrafine permanent marker on the discal cell of the hindwing and were fed ad libitum using untreated 20% honey-water solution on small dish sponges. Twice per day, at 09:00 h and 16:00 h, mating pairs and deaths were recorded. Each cage was also observed for a total of eight 10-min intervals (times dispersed throughout the day between 09:00–16:00) to record the duration of flying, mating and feeding for each individual monarch. After 5 days, all butterflies were weighed immediately following removal.
To measure reproduction, we placed 4 mated females per treatment (10 treatments, n = 40 females in total) into individual oviposition cages (1 monarch/ milkweed stalk) for 3 days with 20% honey water on sponges provided ad libitum. The total number of eggs laid per female at the end of the 3-day interval was recorded and the proportion of eggs hatched after 5 days was quantified to the nearest 10%. All other butterflies were placed into an incubator at 12°C and held to record the time in days until death to determine relative longevity without feeding per treatment.
Experiment 1 analyses
Analyses for both experiments were performed in R version 3.5.3. We first used linear mixed models (normal error structures) in the package lme4 (Bates, Maechler, Bolker, & Walker, 2015) to test for relationships between neonicotinoid treatment and several weight metrics: mean and total nectar consumption (post- minus pre- feeding weight), weight before and after entering the flight cages and the difference in weight between eclosion and exiting the mating cages, using the following model structures: [response variable = insecticide type * dose + sex + genetic lineage (random effect) + weight at eclosion (only for consumption variables)]. The number of eggs laid, the number of times male monarchs mated, the total number of days monarchs survived and the proportion of eggs hatched were analysed using generalised linear mixed models in the package afex (Singmann, Bolker, Westfall, Aust, & Ben-Shachar, 2019) which wraps around the lme4 package (Bates et al., 2015) with Poisson error distributions, using the same model above, with eclosion weight included as a continuous covariate. Residuals were checked for normality and equal variances where appropriate. All response variables were tested using separate generalised and linear models. A principal component analysis was used to reduce the suite of behavioural observations (time spent flying, mating, or feeding) into a single response variable; 36.89% of the variation in the data was explained by PC-1, with variable loadings as follows: 0.369 for feeding, 0.332 for flying and 0.299 for mating. We examined how the first principal component (PC-1) depended on insecticide treatments (type * dose) and monarch sex.
Experiment 2: effects of high dose clothianidin on survival, weight and flight of lab-reared and wild monarchs
We exposed 40 wild-caught fall migrant monarchs that overwintered in the lab and 12 captive-reared monarchs, to clothianidin at 0, 909 and 4030 ng/mL using 20% honey-water solutions to test effects on survival, weight and flight performance. Monarchs were fed every second day for 10 days. We weighed monarchs before and after each feeding to the nearest 0.001 g. Adults were stored in individual glassine envelopes at 23°C (14 h daylength) in the same incubator between feedings.
After the fifth and final feeding, we flew monarchs on a near-frictionless tethered flight mill to measure flight speed and distance (Figure S3). Flight trials were conducted indoors during March 2020 in a 9 m2 room at 29.7°C (range 27.8–31.4°C) between 10:00 h and 17:30 h. On the last feeding day, we glued lightweight steel wires to the dorsal thorax of each monarch using rubber cement, following Bradley and Altizer (2005). As per Schroeder, Majewska, and Altizer (2020), the average weight of the wire attachment was 0.19 g (range 0.10–0.33 g). Monarchs were placed into 0.6 m3 mesh cages to adjust to the weight of the wire for 12–24 h, with 20% (v/v) honey water provided ad libitum. The flight mill was constructed as described in Bradley and Altizer (2005) and Fritzsche McKay, Ezenwa, and Altizer (2016) from a 120 cm lightweight carbon rod with a diameter of 3 mm (4.23 m circumference) attached to a nearly frictionless steel pivot (Figure S3). We tethered monarchs to one end of the horizontal rod and a flag at the opposite end passed through an infrared beam on a photo-gate to estimate flight velocity per revolution (m/s; software PASCO Capstone). Windows were covered with white paper to prevent monarchs from responding to sun angle cues during flight and we set four-floor lamps to provide an even distribution of light.
Monarchs were flown for a maximum of 1 h. If monarchs stopped flying for more than 5 s, they were agitated with a gust of air for up to three times. If the monarch did not resume flight after three agitations, the flight was terminated and monarchs were returned to mesh cages. For each flight, we calculated the distance (km) flown and the total time in flight (hr). We calculated flight velocity by dividing the circumference of the flight path by the time to completion of a revolution. Average velocity across the entire trial was calculated as the mean of all velocities per revolution.
Roughly 2 h after monarchs were flown, we performed a drop test to quantify the reaction time of each butterfly. Monarchs were grasped with 2 fingers holding all wings together close to the thorax and dropped from approximately 1.5 m to the floor. The time (s) taken to open their wings was measured by stopwatch; monarchs that did not open their wings before landing on the ground were recorded with drop time as one full second. Monarchs that successfully opened their wings did not exceed 0.8 s.
Experiment 2 analyses
We analysed average and total nectar consumption and flight speed and distance using linear models (normal error structures), with clothianidin dose (0, 909, or 4030 ng/mL), monarch sex and monarch source (wild or reared) as predictor variables. We used a generalised linear model (GLM) with a Poisson error structure to analyse adult longevity in days. Model residuals were checked for normality and equal variances where appropriate. For flight variables, we included weight before first feeding and weight change between the first and last feeding as continuous covariates. Analyses and data for both experiments are available at https://github.com/codyprouty/Adult-Neonicotinoids.
RESULTS
Neonicotinoid use and residue analysis
Residue samples from Experiment 1 were close to the intended applied dose, with the lowest measuring 54% below intended dose and the highest measuring 6.2% above the intended dose (Table 1). In Experiment 2, samples were within 80%–90% of the intended applied dosage, with one sample at 909 ng/mL instead of 1000 ng/mL and another sample at 4030 ng/mL instead of 5000 ng/mL (Table 1). Because residue metrics differed from the intended applied doses, we used the HPLC-generated values as independent variables in all analyses.
| Neonicotinoid type | Intended dose (ng/mL) | Realised dose (ng/mL) |
|---|---|---|
| Clothianidin | 25 | 19 |
| Imidacloprid | 25 | 15 |
| Clothianidin | 50 | 36 |
| Imidacloprid | 50 | 28 |
| Clothianidin | 100 | 75 |
| Imidacloprid | 100 | 54 |
| Clothianidin | 500 | 531 |
| Imidacloprid | 500 | 386 |
| Clothianidin | 1000 | 909 |
| Clothianidin | 5000 | 4030 |
Wild-caught bumble bee bioassay to confirm spiked solution dose
After 4 h, all four B. impatiens in the 0 ng/mL (honey water only) treatment remained actively flying and had to be chilled at 14°C prior to removal. For imidacloprid, the bees treated with 15 ng/mL were immobile for 4 h but still alive; agitating the container did not cause flight. At 28 and 50 ng/mL of imidacloprid, bees were also lethargic and were immobile by 4 h, but still alive. At the 386 ng/mL imidacloprid, all bees died within 2 h of exposure. For clothianidin, all bees from both the 19 and 36 ng/mL treatments were lethargic and moving slower than controls and unable to fly by 4 h. For the 75 ng/mL treatment of clothianidin, 50% of bees were dead by 4 h and the rest were slow and lethargic. For the 531 ng/mL treatment, all bees died after 2 h. Collectively, these findings demonstrate toxicity of the neonicotinoid treatments for the bee species tested here.
Experiment 1: low dose neonicotinoid effects on monarch movement and reproduction
The vast majority (97.8%, N = 316) of adult monarchs survived throughout the feeding treatment period. The 7 dead monarchs were distributed across feeding treatments, including controls. The proportional change in monarch weight over the course of the experiment ((final–initial)/initial) did not differ among insecticide treatments (Figure 1a, Table 2) and females gained proportionately more weight than males (Table 2).
| Response variable | Predictors | Estimate | Df | t/z value | p value |
|---|---|---|---|---|---|
| Weight post cages | Intercept | 1.56 × 10−1 | 7.85 | <0.001 | |
| Dose | 1.09 × 10−4 | 1 | 1.29 | 0.199 | |
| Neonicotinoid type | 2.22 × 10−2 | 1 | 0.97 | 0.335 | |
| Sex | −3.16 × 10−1 | 1 | −16.11 | <0.001 | |
| Dose: neonic | −3.17 × 10−5 | 1 | −0.23 | 0.820 | |
| Final-initial weight | Intercept | 9.09 × 10−2 | 7.21 | <0.001 | |
| Dose | 7.35 × 10−5 | 1 | 1.36 | 0.176 | |
| Neonicotinoid type | 1.61 × 10−2 | 1 | 1.09 | 0.275 | |
| Sex | −1.99 × 10−1 | 1 | −15.83 | <0.001 | |
| Dose: neonic | −2.63 × 10−5 | 1 | −0.29 | 0.768 | |
| Average nectar consumed | Intercept | 9.36 × 10−2 | 4.46 | <0.001 | |
| Dose | 2.30 × 10−5 | 1 | 1.34 | 0.182 | |
| Neonicotinoid type | −6.09 × 10−3 | 1 | −1.33 | 0.185 | |
| Sex | −8.47 × 10−3 | 1 | −1.96 | 0.051 | |
| Weight at eclosion | 1.09 × 10−1 | 1 | 3.63 | <0.001 | |
| Dose: neonic | 3.52 × 10−5 | 1 | 1.25 | 0.212 | |
| Proportion of Eggs hatched | Intercept | 1.31 | 1.11 | 0.266 | |
| Dose | −5.07 × 10−4 | 1 | −0.57 | 0.572 | |
| Neonicotinoid type | 5.71 × 10−2 | 1 | 0.43 | 0.667 | |
| Weight at eclosion | −9.73 × 10−1 | 1 | −0.51 | 0.610 | |
| Dose: neonic | −7.02 × 10−5 | 1 | −0.08 | 0.939 | |
| Days survived in envelopes | Intercept | 3.10 | 29.91 | <0.001 | |
| Dose | −3.16 × 10−5 | 1 | −0.41 | 0.682 | |
| Neonicotinoid type | 3.82 × 10−3 | 1 | 0.30 | 0.761 | |
| Sex | −3.16 × 10−1 | 1 | −13.15 | <0.001 | |
| Weight at eclosion | 5.94 × 10−1 | 1 | 3.57 | <0.001 | |
| Dose: neonic | −4.72 × 10−5 | 1 | −0.62 | 0.535 | |
| Times males mated | Intercept | 1.48 | 2.50 | 0.012 | |
| Dose | −1.58 × 10−3 | 1 | −2.51 | 0.012 | |
| Neonicotinoid type | 6.02 × 10−2 | 1 | 0.89 | 0.376 | |
| Weight at eclosion | −1.12 | 1 | −1.27 | 0.204 | |
| dose: neonic | 1.21 × 10−4 | 1 | 0.19 | 0.847 | |
| Number of eggs laid | Intercept | 4.13 | 26.79 | <0.001 | |
| Dose | −1.76 × 10−3 | 1 | −14.47 | <0.001 | |
| Neonicotinoid type | 4.07 × 10−2 | 1 | 2.90 | 0.004 | |
| Weight at eclosion | 1.89 | 1 | 8.66 | <0.001 | |
| Dose: neonic | 1.16 × 10−3 | 1 | 9.14 | <0.001 | |
| Principal component 1 | Intercept | 2.99 × 10−2 | 0.25 | 0.804 | |
| Dose | 1.06 × 10−4 | 1 | 0.20 | 0.842 | |
| Neonicotinoid type | −7.24 × 10−2 | 1 | −0.51 | 0.611 | |
| Sex | −2.80 × 10−2 | 1 | −0.23 | 0.819 | |
| Dose: neonic | 2.98 × 10−4 | 1 | 0.34 | 0.732 |
- Note: Effects of low-dose neonicotinoids on weight and reproduction of lab-reared monarchs. Variables that were analysed using normal error structures included weight following removal from flight cages, proportional weight change, average nectar consumed and the first principal component from an analysis of mating, flying and feeding behaviours. Count data, including the proportion of eggs hatched, number of days survived, number of times mated and number of eggs laid were analysed using Poisson error structures. Full models are described in the Methods text. Rows in bold indicate significant effects.
The number of times male monarchs mated significantly decreased with higher doses of both neonicotinoids (Figure 1b; Table 2). At 0 ng/monarch, male monarchs mated on average 2.08 (SE: 0.15) times and at 500 ng/monarch, mated 0.945 (SE: 0.27) times as calculated by estimated marginal means. The number of eggs laid by females decreased with higher imidacloprid (386 ng/mL), but not clothianidin doses (Figure 1c). This interaction between neonicotinoid type and dose was significant (Table 2), where only higher concentrations of imidacloprid led to reduction in the number of eggs laid and clothianidin had no significant effect on the number of eggs laid. For imidacloprid, monarchs fed 0 ng/monarch laid 167 (SE: 12.9) eggs on average and those fed 500 ng/monarch laid 44.1 (SE: 5.78) eggs as calculated by estimated marginal means. The proportion of eggs that hatched did not differ in response to neonicotinoid type or dose (Table 2). Time monarchs spent mating, flying and feeding as summarised in PC-1, showed no relationship with neonicotinoid treatment (Table 2; Figure S4).
Experiment 2: high dose clothianidin effects on survival, weigh, and flight
A total of 69.7% of monarchs (46/67) at the start of the experiment survived the 10-day feeding period. Monarchs treated with both clothianidin doses had significantly reduced survival: 95.5% of controls survived compared to 58.8% at 909 ng/mL and 57.8% at 4030 ng/mL (Figure 2; Table 3). The average volume of nectar monarchs consumed was significantly lower when treated with 4030 ng/mL of clothianidin, with 0.059 g consumed compared to 0.102 g in controls (Figure 3a, Table 3). Male monarchs lost more weight than females and monarchs that weighed more prior to feeding lost more weight during the 10-day feeding period (Table 3). Before feeding, adult monarch weight did not differ significantly between the treatments (F1,62 = 0.045, p = 0.833).
Across all treatments (n = 35), monarchs flew an average distance of 1.77 km ± 0.22 SE at a speed of 2.95 km/h ± 0.15 SE. Monarchs treated with clothianidin flew for shorter distances at slower speeds (Figure 3c,d, Table 3), with no effect from any variable other than dose. The drop test showed that monarchs treated with higher clothianidin doses had slower reaction times (were more delayed in opening their wings; Figure 3b, Table 3). The source of monarchs (wild or lab-reared) significantly affected monarchs' proportional weight change during flight, but no other variables depended on monarch source (Table 3). Flight data were excluded from 2 controls (wild) and 9 treatments (2 lab-reared, 7 wild) monarchs that did not fly for at least 2 min.
| Response variable | Predictors | Estimate | Df | t/z value | p value |
|---|---|---|---|---|---|
| Average nectar consumed | Intercept | 1.52 × 10−1 | 7.92 | <0.001 | |
| Dose (909) | −5.75 × 10−3 | 1 | −0.58 | 0.564 | |
| Dose (4030) | −4.33 × 10−2 | 1 | −4.55 | <0.001 | |
| Monarch source | 6.54 × 10−3 | 1 | 0.63 | 0.530 | |
| Sex | −1.90 × 10−2 | 1 | −2.30 | 0.025 | |
| Weight pre-exposure | −7.56 × 10−2 | 1 | −2.56 | 0.013 | |
| Proportional weight change during feeding | Intercept | −1.97 × 10−3 | −0.04 | 0.968 | |
| Dose (909) | 8.25 × 10−2 | 1 | 1.89 | 0.065 | |
| Dose (4030) | 6.90 × 10−2 | 1 | 1.64 | 0.110 | |
| Monarch source | −1.26 × 10−1 | 1 | −2.74 | 0.009 | |
| Sex | −9 × 10−2 | 1 | −2.56 | 0.014 | |
| Distance flown | Intercept | 1.51 | 2.77 | 0.010 | |
| Dose (909) | −1.09 | 1 | −2.51 | 0.018 | |
| Dose (4030) | −1.90 | 1 | −3.17 | 0.004 | |
| Monarch source | 5.74 × 10−1 | 1 | 1.16 | 0.256 | |
| Sex | -1.58 × 10−1 | 1 | −0.41 | 0.683 | |
| Weight change during flight | 3.43 × 101 | 1 | 2.06 | 0.049 | |
| Average flight speed | Intercept | 3.10 | 8.12 | <0.001 | |
| Dose (909) | 2.50 × 10−1 | 1 | 0.82 | 0.419 | |
| Dose (4030) | −1.70 | 1 | −4.07 | <0.001 | |
| Monarch source | 2.08 × 10−1 | 1 | 0.60 | 0.552 | |
| Sex | -1.71 × 10−1 | 1 | −0.64 | 0.526 | |
| Weight change during flight | −1.01 × 101 | 1 | −0.87 | 0.393 | |
| Drop test | Intercept | 3.45 × 10−1 | 2.59 | 0.014 | |
| Dose (909) | 1.89 × 10−1 | 1 | 2.62 | 0.013 | |
| Dose (4030) | 5.92 × 10−1 | 1 | 9.14 | <0.001 | |
| Monarch source | −9.62 × 10−2 | 1 | −1.37 | 0.181 | |
| Sex | 4.38 × 10−3 | 1 | 0.07 | 0.942 | |
| Weight pre-exposure | 7.89 × 10−2 | 1 | 0.35 | 0.729 | |
| Survival | Intercept | −2.42 | −1.22 | 0.222 | |
| Dose (909) | −2.69 | 1 | −2.78 | 0.005 | |
| Dose (4030) | −2.73 | 1 | −2.91 | 0.004 | |
| Monarch source | −5.62 × 10−1 | 1 | −0.84 | 0.400 | |
| Sex | 9.97 | 1 | 2.86 | 0.004 | |
| Weight pre-exposure | 7.38 × 10−2 | 1 | 0.09 | 0.930 |
- Note: Effects of high doses of clothianidin on survival, weight and flight of lab-reared and wild monarchs. All variables except survival (binomial) were analysed using normal error structures. Full models are described in the Methods section. Rows in bold indicate significant effects at α < 0.05.
DISCUSSION
Past research demonstrated that exposure to low doses of neonicotinoids commonly encountered in the environment reduce the foraging behaviour, flight and survival of both native and managed bees. Understandably, the potential risks of neonicotinoid insecticides on monarch butterflies throughout their North American migratory range were raised as a concern for the future persistence of migratory populations (e.g., Lu, Hung, & Cheng, 2020; Stenoien et al., 2018; Thogmartin et al., 2017; Tracy et al., 2019). Results presented here demonstrate adult monarchs tolerate ingested imidacloprid and clothianidin (15–531 ng/mL) beyond the range of reported field residue in nectar (clothianidin: up to 40.7 ng/g in nectar; imidacloprid: up to 72.8 ng/g in nectar; Sanchez-Bayo & Goka, 2014), without apparent lethal or sublethal effects. Higher doses (386 and 531 ng/mL) negatively affected male mating success and female oviposition. Extremely high doses of clothianidin (>900 ng/mL) negatively affected monarch survival, reaction time and flight. Verification that the observed lack of effects was driven by species biology was confirmed by feeding the same neonicotinoid solutions administered to monarchs to field-caught bumblebees, Bombus impatiens (Experiment 1, 15–531 ng/mL clothianidin and imidacloprid), which dramatically reduced bumble bee survival within a period of <4 h.
This study is the first to experimentally show negative reproductive effects on butterflies as a result of adult-stage exposure to higher-than-field residues of imidacloprid and clothianidin. In particular, female monarchs showed reduced fecundity at moderately high doses of imidacloprid compared to clothianidin (386 ng/mL, Experiment 1). Bees and butterflies have major differences in their life histories and behaviour but some similarities may exist in their reproductive responses to neonicotinoids. Bumblebees were shown to have reduced fecundity following low-dose exposure to imidacloprid and at higher doses, bees' ovaries did not develop at all (Laycock et al., 2012), owing to their lower food consumption following exposure to imidacloprid. It is unlikely that reduced fecundity in monarchs was a result of underfeeding, as we observed similar weight changes across exposure treatments. Male monarchs that were fed moderately high (386–531 ng/mL) doses of clothianidin and imidacloprid were shown to mate with females less often. In bees, numerous studies showed reduction in movement (Scholer & Krischik, 2014; Tosi & Nieh, 2017), foraging (Colin, Meikle, Wu, & Barron, 2019; Feltham et al., 2014; Gill & Raine, 2014; Lämsä et al., 2018; Stanley & Raine, 2016), flight (Kenna et al., 2019) and mating (Forfert et al., 2017, Straub et al. 2021). Reductions in male monarch mating success and female oviposition activity could result from reduced neurological functions from the treatment, as these behaviours are complex and energetically demanding (Oberhauser, 1989, Oberhauser & Frey, 1999, Solensky, 2004).
Because North American monarchs undertake an annual long-distance migration of up to 5000 km, even small reductions in activity levels and flight performance could lower migratory success (Reppert & de Roode, 2018). In our results, we showed that high clothianidin exposure (909 and 4030 ng/mL) lowered monarch flight performance (distance and speed) and reaction time (drop test). Wild and lab-reared monarchs showed comparable responses to exposure. Importantly, exposure to field-relevant lower doses did not reduce monarch feeding or flight activity in comparison to other caged monarchs and the higher doses that reduced flight are unlikely to be encountered by monarchs in the wild.
Past studies of adult monarch consumption of neonicotinoids yielded contrasting results. The first, Krischik et al. (2015), found no reductions in survival or fecundity in monarch and painted lady butterflies when force fed 15 or 30 ng/mL imidacloprid in 30% sugar water for 30 days, or when feeding on intact flowers from tropical milkweed plants treated with a soil-applied label rate (1× and 2×) of imidacloprid that resulted in a residue of 6000 ng/mL imidacloprid in flowers. A second study (James, 2019) did find evidence for significant reductions in adult monarch survival at 23.5 ppb (ng/mL) of imidacloprid mixed into 5% sugar water (fed ad libitum for up to 22 days), with no effect on oogenesis. In a more recent study, Krishnan, Jurenka, and Bradbury (2021), performed 96-h bioassays with adult monarchs at 140 ng/mL clothianidin, 250 ng/mL imidacloprid and 330 ng/mL thiamethoxam, which showed no effects on survival. The doses used were at least 100-fold higher than the highest dose measured in nectar of wildflowers adjoining seed-treated fields (Botías et al., 2015; 2016). In our experiments, we found that monarchs did not begin to show negative survival effects until they were treated with at least 909 ng/mL of clothianidin and concentrations lower than this (15, 19, 28, 36, 54, 75, 386, or 531 ng/mL clothianidin and imidacloprid) showed no significant effects on survival. This directly contradicts the findings by James (2019) but supports findings from both Krischik et al. (2015) and Krishnan, Zhang, et al. (2021).
Neonicotinoids, especially clothianidin, imidacloprid and thiamethoxam were shown to affect many aspects of bee foraging and movement in numerous research publications. Tosi et al. (2017) found that honey bees had reduced flight duration, distance and velocity caused by chronic 2-day exposure of the higher dose of the neonicotinoid thiamethoxam at doses of 1.96–2.90 ng/bee/day. In our experiments, it required much higher levels of clothianidin and imidacloprid to reduce flight performance. Homing flights following catch and release experiments were also shown to be greatly reduced in honeybees that were treated with between 25 and 12,500 ng/mL of clothianidin, imidacloprid and thiacloprid (Fischer et al., 2014). In bumblebees, thiamethoxam reduced bees' foraging ability at 10 ng/mL (Stanley & Raine, 2016), imidacloprid led to slower foraging rates and fewer flower visits at 1 ng/mL (Lämsä et al., 2018) and imidacloprid delayed behavioural changes during foraging at 10 ng/mL (Phelps et al., 2020). Experiments that test foraging ability and preferences in monarchs would give a nice comparison to the honey bee and bumble bee literature. In our current set of experiments, we force-fed monarchs insecticides (Figure S1), rather than having them forage naturally as in bee studies, which could affect our results. One of the few studies on effects of imidacloprid on migratory ability was performed with birds. In white-crowned sparrows, Zonotrichia leucophyrs, sublethal exposure to 1.2 or 3.9 mg/kg body weight caused delays in initiating flight at the higher concentration that could cause migratory delays and reduce survival (Eng, Stutchbury, & Morrissey, 2019). Given reductions in bee mobility following insecticide exposure and the known presence of neonicotinoids along agricultural field margins, it seems plausible that chronic pesticide exposure at both larval and adult stages could lower monarch flight performance during the fall migration and could warrant further testing.
Investigating the interactive effects of nectar limitation, exposure to neonicotinoids and other ecological stressors on adult performance metrics should be a priority for future work. For example, honey bees parasitized by Varroa destructor show reduced immune response when exposed to neonicotinoids, leading to increased parasite reproductive success (Annoscia et al., 2020) and reduced flight performance when exposed to both stressors (but not when exposed to each individually (Blanken, van Langevelde, & van Dooremalen, 2015)). Our first experiment reared monarchs under low density, with ample food and ideal temperatures during development. However, our wild-caught monarchs would have experienced a variety of these stressors. An experiment that compares these effects in the field, under cases of food limitation or other sub-optimal conditions, would be an important next step to address whether these effects reported here hold up across a range of environmental circumstances.
AUTHOR CONTRIBUTIONS
CP, VK and SA conceived the ideas and designed methodology; CP collected the data; CP and LB analysed and visualised the data; CP and SA led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.
ACKNOWLEDGEMENTS
We thank Seth Wenger, Richard Hall, Andy Davis, and members of the Altizer and Hall labs at UGA for comments on project design and earlier drafts of the manuscript. We thank Ashley Ballew, Farran Smith, Kade Donaldson, Jonathan Schulz, Jaycee Quinn, Caroline Aikins, Kathleen Clancy and Christopher Brandon for assistance in rearing monarchs and data collection. CP was supported by the Lincoln P. Brower Grant from the Monarch Butterfly Fund. SA was supported by NSF DEB-1754392 and by SERDP-RC2700.
CONFLICT OF INTEREST STATEMENT
The authors declare there are no conflicts of interest.
Open Research
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are openly available in Adult-Neonicotinoids at https://github.com/codyprouty/Adult-Neonicotinoids.
