Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Non-detection of honeybee hive contamination following Vespula wasp baiting with protein containing fipronil

  • Eric D. Edwards ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing (EDE); (RAK)

    Affiliation Department of Conservation, Wellington, New Zealand

  • Ethan F. Woolly,

    Roles Data curation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing

    Affiliation School of Chemical and Physical Sciences, Victoria University of Wellington, Wellington, New Zealand

  • Rose M. McLellan,

    Roles Data curation, Formal analysis, Investigation, Writing – review & editing

    Affiliation School of Chemical and Physical Sciences, Victoria University of Wellington, Wellington, New Zealand

  • Robert A. Keyzers

    Roles Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing (EDE); (RAK)

    Affiliations School of Chemical and Physical Sciences, Victoria University of Wellington, Wellington, New Zealand, Centre for Biodiversity and Restoration Ecology, Victoria University of Wellington, Wellington, New Zealand


Introduced wasps (Vespula germanica and V. vulgaris) are costly invertebrate pests in New Zealand, with large impacts on the local ecology and economy. Wasps eat honeybees (Apis mellifera), which has potentially devastating effects on hive health, as well as agricultural and horticultural industries. Vespex bait, which contains fipronil in a proteinaceous carrier, has recently been introduced for wasp control. In over a decade of reported trials, honeybees have never been observed foraging on Vespex, likely because the bait contains no sugars to serve as a bee food source. However, the potential for the control agent fipronil to enter beehives has not been tested. Therefore, here, we investigated this using a liquid chromatography–mass spectrometry assay of fipronil and two of its environmental breakdown and metabolic derivatives, fipronil desulfinyl and fipronil sulfone. We did not detect fipronil in any of the worker bee, bee larva, honey or pollen samples (n = 120 per product) collected from 30 hives over a 2-year period. Furthermore, although we detected fipronil desulfinyl in one honeybee sample, this is thought to have originated from a single individual, representing a rare occurrence of intoxication, and there was no evidence that Vespex was the toxicant source. There was also no evidence of trophallactic transfer of fipronil or its derivatives in any of the hives sampled. Previous studies have reported the impairment of individual bee performance at fipronil doses similar to the detection limit of our study. However, our results provide confidence that if undetectable intoxication was occurring, it would involve an acute exposure for those few individuals affected, with minimal impairment to colonies. Therefore, we conclude that the use of Vespex in the vicinity of honeybees does not result in significant hive uptake while effectively reducing wasp pressure on honeybee colonies.


Unintentional, non-target exposure during the broad-scale use of insecticides has been recognised as a contributing factor to the loss of honeybees (Apis mellifera L. 1758) worldwide [18]. However, the use of chemical insecticides remains the most effective method for suppressing many insect pests. Therefore, it is important that non-target effects, particularly on native or economically valuable species, are minimised during insect pest interventions.

Two exotic species of wasps have been accidentally introduced to New Zealand: Vespula germanica (Fabricius, 1793) in 1945 and V. vulgaris L., 1758 in the late 1970s [910]. Both species are now distributed throughout New Zealand, but V. vulgaris is most abundant. Their pest status is derived from the harm they bring to native species and ecosystems [1115], as well as the local economy [16]. Vespula wasp attacks on honeybees have been widely reported [14, 1720], and it has been conservatively estimated that New Zealand beekeepers lose NZ$8.8M per year as a result of wasps taking honey and bees from hives, bees defending their hives instead of foraging, and the protection of hives from wasp attack, which includes the use of insecticides [16].

A new method for suppressing wasps over a wide area recently became available in New Zealand [21]. This involves placing bait stations on trees in areas where wasps are active and filling them once per year with a protein wasp bait containing 1 g/kg fipronil active ingredient (traded as Vespex). Since this bait does not contain any sugars, it is improbable that honeybees will visit. However, since hive safety is of paramount importance in any insecticide-based control strategy, the dispersal and potential uptake of fipronil into hives needs to be assessed to ensure that it is safe to use in the field.

Potential for honeybee exposure

One factor that limits fipronil exposure for invertebrates other than Vespula wasps is the lack of sugars in the bait matrix, which consists of a fresh protein paste. When wasps are abundant and are shown to be seeking protein [21], 20–30 g of this paste is placed in bait wells that are inserted into weatherproof bait stations mounted on trees approximately 1.5 m above the ground, making them accessible to winged insects. The potential period of exposure to the bait is brief, as any residual paste dries to a solid within 3 days and is removed within 8 days of initial placement, and treatment is not repeated until the following year.

Worker wasps carrying bait have a limited ability to digest protein and so fly to their nest, where the larvae are able to share any digested food, spreading fipronil active ingredient throughout the wasp colony. These colonies are usually found in dry soils, and any dead wasp larvae, pupae and adults decay in situ and can be scavenged by rodents. Any fipronil residues would therefore be concentrated at wasp nest sites and become bound to organic materials. Bait stations can be located 50 m apart but are usually at a lower density. Based on the recommended dosing regimen, the maximum amount of fipronil and its metabolites that would enter the environment per 10,000 m2 per year is 0.12 g, which would be achieved by wasps taking all of the bait to their nests within 2–3 days. However, this amount is more typically <0.02 g per 10,000 m2.

There is no obvious pathway by which Vespex would have lethal or sub-lethal effects on honeybees. However, there are four theoretical pathways for honeybee exposure to fipronil. First, honeybee workers occasionally enter wasp bait stations, which would bring them in dermal contact with the bait or oral contact while seeking mineral salts [2223]. Second, moribund intoxicated wasps may enter bee hives. Third, residues of fipronil and its environmental degradation and metabolic products may be taken up systemically by plants at the wasp nest site, which may be subsequently visited by bees [2425], although it seems unlikely that this would occur at a scale that could harm bee colonies and so this pathway was not addressed by this study. Finally, insecticide containing fipronil active ingredient is sprayed onto onion crops (Allium spp.: Amaryllidaceae) and stock food brassicas (Brassica spp.: Brassicaceae) in New Zealand, which could represent an alternative environmental source [2627].

Detecting honeybee colony impacts

Many studies that have assessed the impacts of fipronil and other insecticides on honeybee colonies have been laboratory based and have only considered the lethal and sub-lethal effects on individual bees [2830]. Therefore, it can be challenging to transfer the findings to field situations of normal apicultural operations. Where a hive fails or has numerous moribund and dead bees that test positive for a toxic agent (or its metabolites), this can be linked to inappropriate insecticide use. However, hive failure can often occur in the presence of multiple environmental stress factors, with no obvious single cause [8, 31]. In many countries, there is great concern that the additive effects of environmental stress factors, including sub-lethal insecticide exposure, may cause widespread hive losses or reductions in pollination services from hives [1, 8, 3234], and attempts to determine the adverse effects of sub-lethal insecticide exposure have linked the detection of trace amounts of insecticide in the hive with laboratory research results for behavioural and physiological effects [8, 33, 35].

A small number of studies have shown that individual worker bees fail to perform at doses of fipronil and its metabolites that are below the limit of detection (LOD) of contemporary assay methods, particularly in the case of thorax dermal contact with solutions [29, 30, 36, 37]. Some studies have also attempted to trial a limited period of sub-lethal exposure to fipronil in the environment and then measure hive performance and metrics of hive health in the following months [3840]. However, a range of confounding factors could affect these results, particularly where hive health and bee performance monitoring occurs weeks or months later [8].

In New Zealand, the potential for bees to be exposed to fipronil when the wasp bait station method is used is acute (via bait or intoxicated wasps; from dermal or oral contact; and within 1–3 days) rather than chronic repeated doses over a week or more from contaminated plants. Consequently, it would be challenging to design a series of treatments and controls that would isolate a potentially enduring hive impact (or lack thereof) from a brief exposure to an undetectably low amount of fipronil [8].

In this study, we chose to seek direct evidence of any colony-wide exposure by developing a test for fipronil and its derivatives that was sufficiently sensitive to detect sub-lethal quantities of up to one-tenth of a conservatively reported median lethal dose (LD50) and then testing numerous samples from multiple hives situated at sites where wasps were baited using Vespex with fipronil as the active ingredient. We limited our data collection to the trace detection of fipronil and its derivatives, and sampled worker bees, bee larvae and other bee products. However, we regarded nurse and forager bees to be one of the best indicators of colony intoxication at the time of sampling as they are centrally involved in both the distribution and redistribution of all foods [41].

Chromatographic methods have been proven to be the premier approach for the qualitative and quantitative analysis of metabolites, herbicides and insecticides [4244]. Therefore, we adapted a method for processing bees and a variety of bee products that utilised a QuEChERS (Quick, Easy, Cheap, Effective, Rugged, Safe) protocol [44] followed by liquid chromatography–tandem mass spectrometry (LC-MS/MS) analysis with a quadrupole-time of flight (Q-TOF) mass spectrometer. We initially validated the use of this method for the detection of sub-lethal levels of fipronil and two of its environmental and metabolic breakdown derivatives, fipronil desulfinyl and fipronil sulfone. We then used it to analyse a total of 480 samples of worker bees, bee larvae, honey and pollen (n = 120 per product; Table 1) over 2 consecutive years to assess the uptake of fipronil following its in situ application for wasp control.

Table 1. Timeline of honeybee hive sampling and wasp treatment.

We acknowledge that a lack of detection would not mean that individual bees were not harmed and that whole colonies could be harmed by sub-lethal quantities that remain below the detection limit for all samples. However, we considered this unlikely given the brief period of protein bait exposure (where the bait has a high concentration, 1 mg/g) and the study design used.

Materials and methods

Permission for field investigative work with pest wasps and honeybees was granted by New Zealand Department of Conservation to EDE as an employee of Department.

Study sites

We sampled honeybee hives in two regions in the South Island of New Zealand that were >75 km apart: Big Bush Conservation Area (BB; BB; 41°17′51″S, 173°34′46″E WGS 84), which includes extensive native beech (Fuscospora spp.) forest and some tall Myrtaceae-dominated shrub-land; and Pelorus Bridge Scenic Reserve (PB; 41°47′08″S, 172°49′09″E WGS 84), which includes beech and hardwood forest. Both sites include agricultural grassland and herb-field areas. These sites were selected because beekeepers place bee hives in beech forest areas on a seasonal basis to allow the bees to collect honeydew, which is produced on beech trees by native scale insects (Ultracoelostoma spp.: Coelostomidiidae). However, pest wasps (Vespula spp.) also reach high abundances on the same carbohydrate resource [11].

At BB, a group of 10 hives was surrounded by 9 wasp bait stations placed on trees 1.5 m above the ground at a distance of 20–120 m from the hives, while a second group of 10 non-treatment control hives was placed 1,900 m from the nearest wasp bait station. At PB, a group of 10 hives was placed at the boundary between forest and pasture, and 325 wasp bait stations were placed at 50-m intervals along transects that were 100 m apart along the same forest boundary and inside the forest but not in the pasture, with a distance of 20 m between the hives and the nearest bait stations. These wasp bait stations were distributed over 2.17 km2 of forest, giving a lower density but broader area of coverage than at BB. A second group of 10 non-treatment hives was located 4,500 m from the nearest wasp bait station.

Wasp control

The experiment was repeated over two seasons at BB (February 2015 and 2016) but was only carried out in one season at PB (February 2016) (Table 1). Starting in February, we examined wasp activity on protein foods following the Vespex manufacturer’s instructions. We baited 20 trays placed 5 m apart with approximately 15 g non-toxic fish bait and left them for 1 hour, following which we conducted an instantaneous wasp count. We planned a bait station wasp control operation when the mean wasp count was ≥1 wasp per tray over 20 trays. As soon as at least 3 days of fine weather was forecast (no rain and locally warm air temperatures), bait stations were filled with 20–30 g Vespex wasp bait (a green protein paste without sugars but containing 1 g/kg fipronil). Since any residual bait paste dries to a hard solid within 3 days, any bait remaining in the bait stations was removed after 5 days and according to standard practice was not replaced.

Honeybee hive sampling

To test whether fipronil was already present in the environment, sampling was undertaken prior to wasp control and in nearby, non-treatment hives. Groups of 10 hives were acclimated to each site at least 10 days prior to the first sampling event. During sampling, a professional beekeeper inspected the queen honeybee and collected a minimum of 20 g of worker bees, 2 g of bee larvae from uncapped cells, 2 g of honey from uncapped cells and pollen. In 2015, scrapings of pollen, propolis and wax were collected from each hive, while in 2016, pollen traps were installed in the base of the hives 2–4 days prior to sampling. Tools were cleaned or replaced prior to sampling each hive.

The four products were assayed separately, giving a total of 160 samples per season per area (four hive products from 10 treatment and 10 non-treatment hives across two sampling periods [before and after treatment]). Sampling was carried out twice at BB and once at PB (Table 1), giving a total of 480 samples across the entire study (n = 120 per product). The samples were couriered overnight on ice to an analytical laboratory (School of Chemical and Physical Sciences, Victoria University of Wellington, New Zealand) and stored at −20°C until analysis, which was carried out within 4 months of receipt. Few studies have examined the metabolic fate of fipronil in insect species, but results indicate that in vivo metabolism primarily converts this toxicant into the sulfone form and that most fipronil is recovered in the parent form 24 hours after exposure [45]. In addition, both the desulfinyl and sulfone forms can be produced photochemically [4647]. Therefore, to maximise the likelihood of detecting any fipronil intoxication and transfer within the hive, post-treatment samples were taken within 1–3 days of wasp treatment.

Sample preparation

All solvents (acetonitrile and hexane) that were used for sample preparation and analysis were of analytical grade and were purchased from Fisher Scientific (Suwanee, GA, USA). Water was purified with an Arium 611UV system (Sartorius AG, Germany). The QuEChERS kits, including ceramic homogenisers, extraction and purification salts, and primary secondary amine (PSA), were obtained from Agilent Technologies, Inc. (Folsom, CA, USA). Fipronil (98% purity) was purchased from AK Scientific (Union City, CA, USA), while its two environmental and metabolic derivatives (fipronil sulfone and fipronil desulfinyl) and the internal standard (IS; thiacloprid), which were of 99.9% purity, were purchased from Sigma-Aldrich (St. Louis, MO, USA). Mixed standard solutions were prepared in acetonitrile.

Sample analysis for the presence of fipronil and its derivatives followed the methods of Kasiotis et al. [43] and Paradis et al. [44]. Briefly, sample preparation was optimised for a mass of 1 g of worker bees, bee larvae, honey or pollen. The average weight of an individual honeybee was estimated by weighing groups of 50 bees to standardize the sample sizes used. Each sample was transferred to a 15-mL Falcon tube and mixed with ceramic homogenisers and extraction salts (0.5 g of MgSO4, 0.2 g of NaOAc and 0.2 g of PSA) in deionised water (3 mL) and shaken vigorously for 2 minutes. The IS (1 mL, 400 ng/mL in acetonitrile) was then added along with an additional 6 mL of acetonitrile and 2 mL of hexane. The sample dispersion was centrifuged at 4,000 × g for 5 minutes at 15°C, following which the supernatant (7 mL) was recovered, transferred to a new Falcon tube containing purification salts and a stationary phase (1.2 g of MgSO4, 0.4 g of PSA and 0.4 g of C18EC) and vortexed for 2 minutes. This was then filtered through cotton wool into a test tube and the retentate was washed with acetonitrile (3 mL). The solvent was evaporated to dryness and then reconstituted in 3:2 acetonitrile/water (1 mL), which was filtered through a 0.45-μm PTFE disk into 2-mL LC-MS vials for analysis. To establish calibration curves (1–100 ng/mL), commercially available blank honey was analysed in which no fipronil or fipronil derivatives were detected, and this honey was then spiked with pesticide standard solutions and mixed with a vortex mixer to ensure homogenous distribution. All samples and calibration standards were analysed in triplicate, and all analyses were performed blind in the laboratory, with analysts unaware of the source of the samples.

There were some minor differences in sample preparation between 2015 and 2016. In 2015, fipronil desulfinyl was not available as a standard and so was not tested for. In addition, no IS was used in 2015, with sample recoveries instead being monitored by using blank honey spiked with 1 mL of a 200 ng/mL fipronil and fipronil sulfone calibration solution for every 24th sample.

LC-MS/MS analysis

Liquid chromatography analysis was carried out using an Agilent Technologies (Palo Alto, CA, USA) 6530 Q-TOF LC-MS system equipped with a Kinetex C18 column (XB-C18; Phenomenex; 2.6 μm particles, 100Å, 50 × 2.10 mm). The chromatographic gradient was optimised for separation of fipronil and its two derivatives along with the IS (S1 Fig). Chromatographic separation was performed at 25°C with an injected volume of 1 μL. The mobile phase was a mixture of (A) water (5 mM ammonium formate) and (B) acetonitrile (0.1% formic acid) with a flow rate of 0.4 mL/min. The target compounds were separated with the following gradient: 50% (B) from 0 to 0.5 minutes; a linear gradient from 50% to 90% (B) from 0.5 to 60 minutes; 90% (B) from 6 to 7 minutes; a linear gradient from 90% to 50% (B) from 7 to 10 minutes; and 50% (B) from 10 to 12 minutes to return to the initial conditions. Ionisation was performed with an electrospray ionisation source (ESI) heated at 300°C in positive-ion mode for the IS and negative-ion mode for fipronil and its derivatives. The following ion source conditions were used: capillary voltage, 2.75 kV; gas flow, 9 L/min; nebuliser, 30 psi; nozzle voltage, 500 V; fragmentor voltage, 140 V; MS acquisition rate, 2 spectra/s; MS m/z range, 100–1100; MS/MS acquisition rate, 2 spectra/s; MS/MS m/z range, 50–500; collision-induced dissociation (CID) gas, N2; CID gas pressure, 16 psi; isolation width, m/z 4; collision energies, 17 or 20 eV; total analysis time, 10 minutes. Any pesticide signals that were detected were normalised to the IS intensity. Agilent Technology’s MassHunter Workstation software, including Qualitative Analysis and Quantitative Analysis (version B.06.00), was used to process the MS and MS/MS data.

Method validation

The LC-MS/MS method [44] was evaluated using several criteria (linearity, recovery rates, repeatability and specificity; see S1 Table) prior to its implementation for real-world samples. Calibration was performed using authentic mixed standards at 11 different concentrations and was repeated using blank honey samples spiked with the pesticide standard solutions. The linearity of the calibration curves was evaluated by obtaining determination coefficients without forcing the line of best fit through zero. Recovery rates were calculated as the ratio between the measured area counts of the spiked and non-spiked samples. The repeatability, specificity, LOD (signal to noise ratio [SNR] 3:1) and limit of quantitation (LOQ; SNR 10:1) were determined in accordance with the International Conference on Harmonization (ICH) guidelines [48]. Fifty worker honeybees were weighed in triplicate to estimate the average weight per bee, to allow validation of the results relative to individual bees.


Method validation

The samples of 50 worker honeybees weighed 6.143 g, 7.436 g and 6.589 g (mean = 6.723, standard error = 5.64%), providing an estimate of 0.134 g per bee or 7 honeybees per gram.

The calibration curves had very good linearity, with r2 values of >0.99 for the standards and >0.98 for the spiked samples (Table 2). The recovery rates for fipronil and its derivatives were relatively low, ranging between 35% and 44%, with relative standard deviations of <11% indicating a high repeatability of the measurements (S1 Table). Specificity was ensured through the use of a retention time (± 0.5 min) and quantifier:qualifier ion ratios (± 0.25–0.45) to avoid false identifications (Table 2). The LODs for fipronil and its two derivatives ranged between 1.1 and 1.8 ng/mL, corresponding to 0.16–0.26 ng/bee or 1.1–1.8 ng/g of bee product, while the LOQs fell between 1.8–2.0 ng/mL (0.25–0.29 ng/bee), which are within the sub-lethal levels needed to observe toxicological effects on bees and below the most conservative reported ingested LD50 of fipronil for honeybees (1 ng/bee; see S2 Table and [36]).

Table 2. Figures of merit for the LC–MS/MS analysis of fipronil and its derivatives.

Testing for the presence of fipronil and its derivatives in honeybee hives

None of the 160 samples (40 samples each of honey, worker bees, bee larvae and a mix of pollen, wax and propolis) that were examined in 2015 tested positive for fipronil above the LOD. Three honey samples that were collected prior to bait deployment did return trace levels (between the LOD and LOQ) of fipronil sulfone, however, although all of these failed to register above the LOD for any of the compounds of interest when freshly prepared samples were retested.

Among the 320 samples comprising honey, worker bees, bee larvae and pollen that were examined in 2016, only one field sample of worker bees from the wasp treatment area in BB returned a positive result for any of the compounds of interest. This sample tested positive for fipronil desulfinyl (6 ng/mL) at a level that was well above the LOQ and at the typical 24–48-hour LD50 for an individual bee for both oral and dermal routes of exposure (see S2 Table). However, two additional re-assays of bees from the same field sample yielded negative results for this compound, as well as for fipronil and fipronil sulfone.


For fipronil and indicator derivatives, a detection limit of 1.1–1.8 ng/mL converts to 0.16–0.26 ng/bee. This is close to one order of magnitude less than a conservatively reported LD50 for a worker bee (S1 Table). However, a range of adverse behavioural and physiological effects for individuals are reported at and below this level of intoxication [1,28,29,30,36,37,49].

We found no trace of fipronil above its detection limit in any of the worker bee, bee larva, honey or pollen samples collected from either study area across both seasons following Vespex bait treatment for wasp control (Table 2). Furthermore, although we detected 6 ng/mL of fipronil desulfinyl in one sample assay containing approximately seven worker bees that was collected following wasp treatment in BB, two further assays from the same bee sample did not detect any of this compound. This suggests that a single bee was likely responsible for the result in the original sample and that this was a rare event, as detections would have been expected in the follow-up assays (or indeed in assays from other co-located hives) if multiple bees were involved in the same sample. However, this level of fipronil desulfinyl is equivalent to the LD50 for a single bee following dermal contact or direct ingestion of fipronil (see S2 Table). To the best of our knowledge, no toxicological data have been published to date for fipronil desulfinyl, but a wide range of oral toxicity values have been cited in the literature for fipronil and its other common derivative, fipronil sulfone (see S2 Table and [36, 38, 43, 4950]).

We acknowledge that honeybees carrying small amounts of fipronil or its derivatives may have been missed through the random selection of bees during both field and laboratory sampling coupled with the low extraction recoveries obtained. Our choice of laboratory sample size (1 g of honeybees) was based on the published protocol of Kasiotis et al. [43] but contrasts with the 5 g of honey used by Paradis et al. [44]. There are advantages and disadvantages to choosing both smaller and larger sample sizes, as smaller sample sizes will avoid the dilution of toxicant residues by any additional extracted biomass but may result in individual positive samples being missed, while larger sample sizes will sample a wider range of the population but may dilute individual insecticide levels. If the 6 ng/mL of fipronil desulfinyl that was detected originated from one individual honeybee, an increase in sample size to 5 g would have resulted in an 80% dilution in the amount detected to 1.2 ng, which is close to our LOD. In addition, the detection of fipronil desulfinyl in 1 out of 80 samples from two study sites is unlikely to have increased had the sample size been doubled but would have imposed significant constraints in terms of cost and time. Therefore, the protocol was optimised through the use of a sensitive method combined with a limited laboratory sample size. In addition, the risk of under-sampling due to the use of a smaller extraction mass can be offset by testing multiple hives (see Table 1), as carried out in this study.

As mentioned above, the finding that no other samples from the same hive or from 29 other hives in wasp treatment areas contained fipronil desulfinyl suggests that this result may have originated from a single honeybee. Assuming the complete photochemical degradation of fipronil to fipronil desulfinyl (see [47]), 6 ng of fipronil desulfinyl could have originated from approximately 10 μg of Vespex bait, which we posit could, for example, have been inadvertently transferred to a single honeybee via walking on or ingesting the protein bait. Alternatively, it is possible that this resulted from an alternative source of contamination, such as a nearby agricultural crop spraying event, as indicated by the pre-treatment result for fipronil sulfone (see below).

The potential exposure of honeybees to fipronil or its derivatives during wasp control will differ from that experienced when fipronil is used as an active ingredient in insecticides for crop seed treatment, crop sprays or ground treatment for termites [51]. In horticultural applications, fipronil insecticide is often applied either as a dust coated on seeds or in aqueous solutions as a spray [51]. The application of fipronil in Vespex differs from this in two ways. First, the active ingredient is present in fresh protein paste wasp baits that are housed in weatherproof trays at a density of no more than three trays per 10,000 m2 and often less than one tray per 10,000 m2. Second, residual wasp bait that is not taken by pest wasps becomes a dry solid within 3 days and is removed from the bait stations after 3–8 days without replacement. Consequently, the environmental exposure of bee colonies to the toxicant (or its environmental or metabolic breakdown products) and the period over which they could develop sub-lethal effects is shorter than indicated by published toxicity tests [36, 40, 5253], although even brief periods of exposure can result in subtle sub-lethal effects with both immediate and delayed effects.

The pathways by which fipronil can be transferred from wasp bait to the bee hive are also limited. Exposure could result from the direct contact of worker bees with the bait, with the entry of intoxicated Vespula wasps being less likely; however, neither scenario has been observed by the authors or reported by beekeepers who have used Vespex. It should also be noted that worker bees appear to have a negligible role as vehicles of fipronil associated with wasp baiting as we found no evidence for the trophallactic transfer of fipronil or its derivatives.

Three honey samples that were obtained from both study areas before wasp baiting occurred returned trace amounts of fipronil sulfone but not fipronil. The retesting of each sample yielded negative results but we cannot exclude the possibility that trace levels of fipronil sulfone just below the LOD were present. Therefore, we consider that these results indicate the low-level presence of this toxicant. The source of these detections was not identified but several possible scenarios exist. First, cross contamination may have occurred in the laboratory. However, this seems extremely unlikely as mixed standards were used and only 3 out of 480 samples returned positive results for only one fipronil derivative. Long-term stability studies on the soil degradation of fipronil have shown that this toxicant can take almost a year to break down (t1/2 111–350 days; [54]). Therefore, given that honeybees do not consume any protein-based foods except pollen, any detection of fipronil in pre-baiting samples will be from historical applications of fipronil in the vicinity of the target sites. Consequently, it is possible that contamination occurred during field sampling or from contaminated nectar sources that were associated with the use of foliar sprays containing fipronil as an active ingredient adjacent to our test sites and within the flight range of honeybees.

These positive results, together with the low LOD/LOQs obtained, demonstrate that the method used was capable of detecting fipronil and its derivatives in bee and bee-product samples–a series of otherwise negative results could bring into question the validity of the chosen method.

>90% reductions in the activity of wasps from the use of Vespex wasp bait is previously reported for similar sites [21]. Since the impacts of the exotic wasps V. vulgaris and V. germanica on bees and bee services have been estimated to cost the New Zealand economy over NZ$70M annually [16], our finding that Vespex was not detected in bee colonies is important. However, there is ongoing concern about the impacts of Vespula wasps globally [13, 14], including in Australia [55], Argentina [5657] and the islands of Hawai’i [5859], indicating that wide-scale studies on the impact of wasp baits on bee populations are required.

Supporting information

S1 Fig. Extracted ion chromatogram of fipronil (2.36 min), fipronil desulfinyl (2.80 min) and fipronil sulfone (3.31 min).


S2 Fig. High-resolution electrospray ionisation tandem mass spectrometry (HR-ESI-MS/MS) fragmentation of fipronil, fipronil desulfinyl and fipronil sulfone.


S3 Fig. Calibration curves for fipronil and its derivatives.

Values are means ± standard errors.


S1 Table. Calculations of figures of merit for fipronil and its derivatives.


S2 Table. Summary of published fipronil LD50 toxicity data for honeybees (Apis mellifera).



We would like to thank Department of Conservation rangers (led by Nic Joice and Jenny Long [St Arnaud], and Phil Clerke [Picton]) for wasp baiting, Ricki Leahy and Ian Henbrey (beekeepers at St Arnaud and Picton) for collecting the samples, and Ian Vorster for his tireless assistance with maintaining the LC-MS/MS instrumentation. Phil Lester, Bruno Fedrizzi and Amanda Todd are also thanked for providing constructive comments on the manuscript. We are grateful to the anonymous reviewers for providing useful comments and therefore improving our manuscript.


  1. 1. Vidau C, Diogon M, Aufauvre J, Fontbonne R, Viguès B, Brunet J-L, et al. Exposure to sublethal doses of fipronil and thiacloprid highly increases mortality of honeybees previously infected by Nosema ceranae. PLoS ONE 2011; 6:e21550. pmid:21738706
  2. 2. Blacquière T, Smagghe G, van Gestel CAM, Mommaerts V. Neonicotinoids in bees: a review on concentrations, side-effects and risk assessment. Ecotoxicology 2012; 21:973–92. pmid:22350105
  3. 3. Vanbergen AJ, Baude M, Biesmeijer JC, Britton NF, Brown MJF, Bryden J, et al. Threats to an ecosystem service: pressures on pollinators. Front Ecol Environ. 2013; 11:251–9.
  4. 4. Sánchez-Bayo F, Goka K. Pesticide residues and bees–a risk assessment. PLoS ONE 2014; 9:e94482. pmid:24718419
  5. 5. Goulson D, Nicholls E, Botías C, Rotheray EL. Bee declines driven by combined stress from parasites, pesticides, and lack of flowers. Science. 2015; 347:1255957. pmid:25721506
  6. 6. Desneux N, Decourtye A, Delpuech JM. The sublethal effects of pesticides on beneficial arthropods. Annu Rev Entomol. 2007; 52:81–106. pmid:16842032
  7. 7. Giorio C, Safer A, Sánchez-Bayo F, Tapparo A, Lentola A, Girolami V, et al. An update of the Worldwide Integrated Assessment (WIA) on systemic insecticides. Part 1: new molecules, metabolism, fate and transport. Environ Sci Pollut Res. 2017; 1–33. pmid:29105037
  8. 8. Pisa L, Goulson D, Yang EC, Gibbons D, Sánchez-Bayo F, Mitchell E, et al. An update of the Worldwide Integrated Assessment (WIA) on systemic insecticides. Part 2: impacts on organisms and ecosystems. Environ Sci Pollut Res. 2017; 1–49. pmid:29124633
  9. 9. Thomas CR. The European wasp (Vespula germanica Fab.) in New Zealand. New Zealand Department of Scientific and Industrial Research Information Series 27. Wellington: Department of Scientific and Industrial Research; 1960.
  10. 10. Donovan BJ. Occurrence of the common wasp, Vespula vulgaris (L.) (Hymenoptera: Vespidae) in New Zealand. NZ J Zool. 1984; 11:417–27.
  11. 11. Moller H, Tilley J, Thomas BW, Gaze PD. Effect of introduced social wasps on the standing crop of honeydew in New Zealand beech forests. NZ J Zool. 1991; 18:171–9.
  12. 12. Toft RJ, Rees JS. Reducing predation of orb-web spiders by controlling common wasps (Vespula vulgaris) in a New Zealand beech forest. Ecol Entomol. 1998; 23:90–5.
  13. 13. Beggs J. The ecological consequences of social wasps (Vespula spp.) invading an ecosystem that has an abundant carbohydrate resource. Biol Conserv. 2001; 99:17–28.
  14. 14. Beggs JR, Brockerhoff EG, Corley JC, Kenis M, Masciocchi M, Muller F, et al. Ecological effects and management of invasive alien Vespidae. BioControl 2011; 56:505–26.
  15. 15. Lester PJ, Beggs JR, Brown RL, Edwards ED, Groenteman R, Toft RJ, et al. The outlook for control of New Zealand’s most abundant, widespread and damaging invertebrate pests: social wasps. NZ Sci Rev. 2013; 70:56–62.
  16. 16. MacIntyre P, Hellstrom J. An evaluation of the costs of pest wasps (Vespula species) in New Zealand. Wellington: Department of Conservation and Ministry for Primary Industries; 2015.
  17. 17. Spradbery JP. Wasps: an account of the biology and natural history of solitary and social wasps. London: Sidgwick and Jackson; 1973.
  18. 18. Clapperton BK, Alspach PA, Moller H, Matheson AG. The impact of common and German wasps (Hymenoptera: Vespidae) on the New Zealand beekeeping industry. NZ J Zool. 1989; 16:325–32.
  19. 19. Coelho JR, Hoagland J. Load-lifting capacities of three species of yellowjackets (Vespula) foraging on honey-bee corpses. Funct Ecol. 1995; 9:171–4.
  20. 20. Wilson EE, Mullen LM, Holway DA. Life history plasticity magnifies the ecological impact of a social wasp invasion. Proc Natl Acad Sci U S A. 2009; 106:12809–13. pmid:19625616
  21. 21. Edwards E, Toft R, Joice N, Westbrooke I. The efficacy of Vespex wasp bait to control Vespula species (Hymenoptera: Vespidae) in New Zealand. Int J Pest Manag. 2017; 63:266–72.
  22. 22. Butler CG. The choice of drinking water by the honeybee. J Exp Biol. 1940; 17:253–61.
  23. 23. Bonoan RE, Tai TM, Rodriguez MT, Feller L, Daddario SR, Czaja RA, et al. Seasonality of salt foraging in honey bees (Apis mellifera). Ecol Entomol. 2017; 42:195–201.
  24. 24. Aajoud A, Raveton M, Aouadi H, Tissut M, Ravanel P. Uptake and xylem transport of fipronil in sunflower. J Agric Food Chem. 2006; 54:5055–60. pmid:16819916
  25. 25. Bonmartin JM, Giorio C, Girolami V, Goulson D, Kreutzweiser DP, Krupke C, et al. Environmental fate and exposure; neonicotinoids and fipronil. Environ Sci Pollut Res Int. 2015; 22(1):35–67. pmid:25096486
  26. 26. Walker GP, MacDonald FH, Wallace AR. Recent developments in management of diamondback moth in New Zealand. In: Srinivasan R, Shelton AM, Collins HL, editors. Proceedings of the Sixth International Workshop on the Management of the Diamondback Moth and Other Crucifer Insect Pests, 21–25 March 2011, Kasetsart University, Nakhon Pathom, Thailand. Shanhua, Taiwan: The World Vegetable Center; 2011. pp. 234–40.
  27. 27. Simon-Delso N, Amaral-Rogers V, Belzunces LP, Bonmatin JM, Chagnon M, Downs C, et al. Systemic insecticides (neonicotinoids and fipronil): trends, uses, mode of action and metabolites. Environ Sci Poll Res. 2015; 22:5–34. pmid:25233913
  28. 28. El Hassani AK, Dacher M, Gauthier M, Armengaud C. Effects of sublethal doses of fipronil on the behavior of the honeybee (Apis mellifera). Pharmacol Biochem Behav. 2005; 82:30–9. pmid:16102801
  29. 29. Aliouane Y, El Hassani AK, Gary V, Armengaud C, Lambin M, Gauthier M. Subchronic exposure of honeybees to sublethal doses of pesticides: Effects on behaviour. Environ Toxicol Chem. 2009; 28:113–22. pmid:18700810
  30. 30. Bovi TS, Zaluski R, Orsi RO. Toxicity and motor changes in Africanized honey bees (Apis mellifera L.) exposed to fipronil and imidacloprid. An Acad Bras Ciênc. 2018; 90(1):239–45. pmid:29641761
  31. 31. Chauzat M-P, Carpentier P, Martel A-C, Bougeard S, Cougoule N, Porta P, et al. Influence of pesticide residues on honey bee (Hymenoptera: Apidae) colony health in France. Environ Entomol. 2009; 38:514–23. pmid:19508759
  32. 32. Decourtye A, Henry M, Desneux N. Overhaul pesticide testing on bees. Nature. 2013; 497:188.
  33. 33. Aufauvre J, Misme-Aucouturier B, Viguès B, Texier C, Delbac F, Blot N. Transcriptome analyses of the honeybee response to Nosema ceranae and insecticides. PLoS ONE. 2014; 9:e91686. pmid:24646894
  34. 34. Sánchez-Bayo F, Goulson D, Pennacchio F, Nazzi F, Goka K, Desneux N. Are bee diseases linked to pesticides?–A brief review. Environ Int. 2016; 89–90:7–11. pmid:26826357
  35. 35. Nunes-Silva P, Hrncir M, Tasso FGJ, Arruda H, Costa L, Pessim G, et al. Applications of RFID technology on the study of bees. 2018. Insectes Sociaux; 10 p.
  36. 36. Roat TC, Carvalho SM, Nocelli RCF, Silva-Zacarin ECM, Palma MS, Malaspina O. Effects of sublethal dose of fipronil on neuron metabolic activity of Africanized honeybees. Arch Environ Contam Toxicol. 2013; 64:456–66. pmid:23224048
  37. 37. Zaluski R, Kadri SM, Alonso DP, Ribolla PEM, de Oliveira Orsi R. Fipronil promotes motor and behavioural changes in honey bees (Apis mellifera) and affects the development of colonies exposed to sublethal doses. Environ Toxicol Chem. 2015; 34:1062–9. pmid:25703042
  38. 38. Mayer DF, Lunden JD. Field and laboratory tests of the effects of fipronil on adult female bees of Apis mellifera, Megachile rotundata and Nomia melanderi. J Apicult Res. 1999; 38:3–4, 191–7.
  39. 39. Nahar N, Ohtani T. Imidacloprid and Fipronil induced abnormal behavior and disturbed homing of forager honey bees Apis mellifera. J Entomol Zool Stud. 2015; 3:65–72.
  40. 40. Kairo G, Poquet Y, Haji H, Tchamitchian S, Cousin M, Bonnet M, et al. Assessment of the toxic effect of pesticides on honey bee drone fertility using laboratory and semifield approaches: A case study of fipronil. Environ Toxicol Chem. 2017; 36:2345–51. pmid:28224659
  41. 41. Scmickl T, Crailsheim K. Inner nest homeostasis in a changing environment with special emphasis on honey bee brood nursing and pollen supply. Apidologie. 2004: 249–63.
  42. 42. Carabajal MD, Arancibia JA, Escandar GM. A green-analytical chemistry method for agrochemical-residue analysis in vegetables. Microchem J. 2016; 128:34–41.
  43. 43. Kasiotis KM, Anagnostopouos C, Anastasiadou P, Machera K. Pesticide residues in honeybees, honey and bee pollen by LC-MS/MS screening: reported death incidents in honeybees. Sci Total Environ. 2014; 485–486:633–42. pmid:24747255
  44. 44. Paradis D, Bérail G, Bonmartin J-M, Belzunces LP. Sensitive analytical methods for 22 relevant insecticides of 3 chemical families in honey by GC-MS/MS and LC-MS/MS. Anal Bioanal Chem. 2014; 406:621–33. pmid:24253411
  45. 45. Durham EW, Siegfried BD, Scharf ME. In vivo and in vitro metabolism of the European corn borer Ostrinia nubilalis. Pest Manag Sci. 2002; 58:799–804. pmid:12192904
  46. 46. Walse SS, Pennington PL, Scott GI, Ferry JL. The fate of fipronil in modular estuarine mesocosms. J Environ Monit. 2004; 6:58–64. pmid:14737471
  47. 47. Ngim KK, Mabury SA, Crosby DG. Elucidation of fipronil photodegradation pathways. J Agric Food Chem. 2000; 48:4661–5. pmid:11052715
  48. 48. ICH. Validation of analytical procedures: text and methodology Q2 (R1). ICH Harmonized Tripartite Guideline. Geneva: ICH; 2005. Available from
  49. 49. Lunardi JS, Zaluski R, Orsi RO. Evaluation of motor changes and toxicity of insecticides fipronil and imidacloprid in Africanized honey bees (Hymenoptera: Apidae). Sociobiology. 2017; 64:50–6.
  50. 50. Carrillo MP, Bovi TS, Negãro AF, Orsi RO. Influence of agrochemicals fipronil and imidacloprid on the learning behavior of Apis mellifera L. Acta Sci Anim Sci. 2013; 35:431–4.
  51. 51. MacBean C (ed). The pesticide manual: a world compendium. 16th ed. Alton: British Crop Protection Council; 2012. ISBN: 9781901396867. PID: 21126852800001951. Available at Accessed on September 2017.
  52. 52. Renzi MT, Amichot M, Pauron D, Tchamitchian S, Brunet J-L, Kretzschmar A, et al. Chronic toxicity and physiological changes induced in the honeybee by the exposure to fipronil and Bacillus thuringiensis spores alone or combined. Ecotoxicol Environ Saf. 2016; 127:205–13. pmid:26866756
  53. 53. Kairo G, Biron DG, Abdelkader FB, Bonnet M, Tchamitchian S, Cousin M, et al. Nosema ceranae, fipronil and their combination compromise honey bee reproduction via changes in male physiology. Sci Rep. 2017; 7:8556. pmid:28819220
  54. 54. Gunasekara AS, Truong T, Goh KS, Spurlock F, Tjeerdema RS. Environmental fate and toxicology of fipronil. J Pestic Sci. 2007; 32:189–99.
  55. 55. Matthews RW, Goodisman MAD, Austin AD, Bashford R. The introduced English wasp, Vespula vulgaris (l.) (Hymenoptera: Vespidae) newly recorded invading native forests in Tasmania. Aust J Entomol. 2000; 39:177–9.
  56. 56. Farji-Brener AG, Corley JC. Successful invasions of hymenopteran insects into NW Patagonia. Ecologia Austral. 1998; 8:237–49.
  57. 57. Masciocchi M, Beggs JR, Carpenter JM, Corley J. Primer registro de Vespula vulgaris (Hymenoptera: Vespidae) en la Argentina. Revista Sociedad Entomológica Argentina. 2010; 69:267–70.
  58. 58. Howarth FG. Impacts of alien land arthropods and mollusks on native plants and animals in Hawaii. Hawaii’s terrestrial ecosystems: preservation and management. Honolulu: University of Hawaii Press; 1985. pp. 149–179.
  59. 59. Foote D, Hanna C, King C, Spurr EB. Efficacy of fipronil for the suppression of invasive yellowjacket wasps in Hawai’i Volcanoes National Park. Cooperative Studies Unit, Technical report HCSU-028. Hilo: University of Hawai’i; 2011.