The authors have declared that no competing interests exist.
Conceived and designed the experiments: OSR GL MC VF. Performed the experiments: OSR. Analyzed the data: OSR. Contributed reagents/materials/analysis tools: OSR GL MC VF. Wrote the paper: OSR GL MC VF.
In recent years, populations of honey bees and other pollinators have been reported to be in decline worldwide. A number of stressors have been identified as potential contributing factors, including the extensive prophylactic use of neonicotinoid insecticides, which are highly toxic to bees, in agriculture. While multiple routes of exposure to these systemic insecticides have been documented for honey bees, contamination from puddle water has not been investigated. In this study, we used a multi-residue method based on LC-MS/MS to analyze samples of puddle water taken in the field during the planting of treated corn and one month later. If honey bees were to collect and drink water from these puddles, our results showed that they would be exposed to various agricultural pesticides. All water samples collected from corn fields were contaminated with at least one neonicotinoid compound, although most contained more than one systemic insecticide. Concentrations of neonicotinoids were higher in early spring, indicating that emission and drifting of contaminated dust during sowing raises contamination levels of puddles. Although the overall average acute risk of drinking water from puddles was relatively low, concentrations of neonicotinoids ranged from 0.01 to 63 µg/L and were sufficient to potentially elicit a wide array of sublethal effects in individuals and colony alike. Our results also suggest that risk assessment of honey bee water resources underestimates the foragers' exposure and consequently miscalculates the risk. In fact, our data shows that honey bees and native pollinators are facing unprecedented cumulative exposure to these insecticides from combined residues in pollen, nectar and water. These findings not only document the impact of this route of exposure for honey bees, they also have implications for the cultivation of a wide variety of crops for which the extensive use of neonicotinoids is currently promoted.
Pollination is a key ecosystem service for both biodiversity and human welfare. Animal-mediated pollination plays a role in the sexual reproduction process of over 90% of the world's angiosperms, thereby sustaining biodiversity and maintaining the integrity of most terrestrial ecosystems
While bees are by far the most efficient group of insect pollinators, their populations are declining worldwide
Although neonicotinoid insecticides can be applied in various ways (pulverization, soil dressing), in North America, they are mainly used as a seed dressing to protect corn and soybean crops from a broad range of root-feeding and sucking pest species. In fact, virtually every single seed of corn and a third of soybean seeds are coated with these insecticides in the US, totalizing more than 110 million acres of land for 2010
Bees can come into contact with these systemic compounds in a number of ways. Recent studies have demonstrated that planting neonicotinoid-coated seeds with a pneumatic drilling machine releases particulate matter contaminated with the insecticides into the environment
In addition to collecting nectar and pollen, honey bees also forage actively for water. High residue levels of neonicotinoids have been measured in guttation and dew water
This study was initiated after noticing how abundant puddles of water were in corn fields following rainfall and anecdotal observations of honey bees drinking from common puddles of rainwater (albeit not from corn fields). The objectives were to 1) examine whether puddles of water from corn fields are contaminated with neonicotinoid compounds and 2) determine the risk associated with the consumption of this water for honey bees. Considering the extent to which these insecticides are used and their remarkably high toxicity, it is essential to thoroughly understand every potential route by which honey bees can be exposed to them.
No ethics approval was required. We obtained private landowners' permission. Private landowners who granted access in this study wish to remain anonymous and specific GPS coordinates cannot be provided as part of that confidentiality. This study did not involve endangered or protected species.
Sampling was conducted in two neighbouring administrative regions in southern Quebec, Canada. Both regions, Montérégie (45° 37′ 10″ N, 72° 57′ 30″ W) and Estrie (45° 24′ 00″ N, 71° 53′ 03″ W), have historically had high levels of agricultural land use. Montérégie alone produces nearly 60% of the province's corn and soybean crops. Since 2008, close to 100% of corn and over two-thirds of soybean crops have been treated with a neonicotinoid coating. Estrie, on the other hand, produces very little corn and soybean, and its agricultural profile is more evenly distributed among a variety of crops whose seeds are generally untreated with neonicotinoids.
Water samples were obtained from puddles of water that had accumulated on the surface of fields following a day of precipitation. All puddles were located at a maximum distance of 1 km from a commercial apiary, well within a honey bee's flight range. In Montérégie, sampling was limited to puddles in corn fields due to the ubiquity of neonicotinoid seed treatment in this crop. Control water samples were collected from puddles in hay fields and grasslands in Estrie and were located at least 3 km from neonicotinoid-coated crops to limit contamination apart and were sampled only once during this study. On June 5th, 2012, 10 samples of water were collected from coated corn fields as corn sowing was still in progress. On May 22nd, 2013, 30 samples were retrieved during corn plantation, half from coated corn fields and half from hay fields and grasslands. An additional 34 water samples were collected from coated corn fields on June 29th, 2013, a full month after sowing had ended. A total of 74 water samples were collected, 15 from untreated crop fields, and 59 from neonicotinoid-treated corn fields. Samples were obtained by collecting water with 50 ml disposable Falcon tubes and filling 1 L amber-coloured glass bottles. Samples were collected from clear water puddles (no suspended solid matter) and tubes were carefully submerged into the puddles to avoid suspending soil particles and to limit sample contamination. Bottles were sealed with aluminum foil-lined lids and immediately placed in a dark cooler. Bottles were stored at 4°C until extraction for chemical analyses, which were done within one week of receiving. Residue analyses were performed by two governmental ISO 17025 accredited laboratories (MAPAQ, CEAEQ).
Water samples collected during corn sowing were analyzed using a modified version of the QuEChERS method originally described by Anastassiades et al. (2003)
Chemical analyses of water result in concentrations expressed in mass of active ingredient per volume of water. In order to understand the potential exposure for bees, the amount of water a honey bee would consume on a daily basis and thus the amount of pesticide it would ingest must be estimated. The drinking water intake rate used in this risk assessment method is based on direct measurement of the water flux rate of the brown paper wasp (
Chemical analyses of puddle water indicated that honey bees are exposed to various agricultural chemicals through collection and consumption of water. A total of 30 different pesticides and metabolites were found in the 74 puddle water samples, with an average of 3.9±2.6 chemicals detected per sample. In the 15 control water samples (untreated-crop fields), 5 pesticides were identified, with some samples containing all 5 and an average of 2.1±3.8 chemicals per sample, always below the limit of quantification. Of the 5 pesticides detected, 4 were herbicides (atrazine, desethylatrazine, metolachlore and simazine) and 1 was a fungicide (thiabendazole). Since occurrence and concentrations of neonicotinoids were similar in water samples collected from corn fields when corn was still being sown, samples from 2012 (10) and 2013 (15) were pooled together in
Pesticide | Class |
Detection | Samples (N) | % | Concentrations (µg/L) | LOQ |
|||
Min | Max | Mean |
SEM |
||||||
Atrazine | HERB, S | 25 | 25 | 100 | 0.1 | 7189.0 | 312.8 | 1434.6 | 0.1 |
Thiabendazole | FUNG, S | 25 | 25 | 100 | 0.1 | 5.7 | 0.6 | 1.3 | 0.1 |
Clothianidin | NEO, S | 23 | 25 | 92 | 0.1 | 55.7 | 4.6 | 12.1 | 0.1 |
Desethylatrazin | HERB | 21 | 25 | 84 | 0.1 | 705.0 | 39.5 | 152.9 | 0.1 |
Thiamethoxam | NEO, S | 18 | 25 | 72 | 0.1 | 63.4 | 7.7 | 16.7 | 0.1 |
Metolachlor | HERB, PS | 11 | 25 | 44 | 0.2 | 10660.0 | 1401.9 | 3353.9 | 0.1 |
Metalaxyl | FUNG, S | 10 | 25 | 40 | 0.1 | 0.7 | 0.4 | 0.2 | 0.1 |
Propazine | HERB | 7 | 25 | 28 | 0.4 | 170.7 | 25.1 | 64.2 | 0.1 |
Spiroxamine | FUNG | 5 | 25 | 20 | 0.4 | 49.5 | 13.9 | 20.1 | 0.1 |
Mesotrione | HERB | 4 | 25 | 16 | 9.7 | 10681.0 | 3437.6 | 5036.5 | 0.1 |
Imazethapyr | HERB | 3 | 25 | 12 | 0.1 | 1.6 | 0.6 | 0.8 | 0.1 |
Boscalid | FUNG, S | 2 | 25 | 8 | 0.2 | 0.8 | 0.5 | 0.4 | 0.1 |
Dimetachlore | HERB | 2 | 25 | 8 | 3.5 | 7.1 | 5.3 | 2.5 | 0.1 |
Dimethenamid | HERB | 2 | 25 | 8 | 0.1 | 0.1 | 0.1 | 0.0 | 0.1 |
Simazine | HERB, S | 2 | 25 | 8 | 1.3 | 40.7 | 21.0 | 27.9 | 0.1 |
Benoxacor | HEBR | 1 | 25 | 4 | 6.1 | 6.1 | 6.1 | NA | 0.1 |
Bentazone | HERB | 1 | 25 | 4 | 1.5 | 1.5 | 1.5 | NA | 0.1 |
Chlorimuron-ethyle | HERB | 1 | 25 | 4 | 0.4 | 0.4 | 0.4 | NA | 0.1 |
Metobromuron | HERB | 1 | 25 | 4 | 1.5 | 1.5 | 1.5 | NA | 0.1 |
Nicosulfuron | HERB, S | 1 | 25 | 4 | 8.4 | 8.4 | 8.4 | NA | 0.1 |
Picoxystrobin | FUNG | 1 | 25 | 4 | 2.5 | 2.5 | 2.5 | NA | 0.1 |
Rimsulfuron | HERB | 1 | 25 | 4 | 6.0 | 6.0 | 6.0 | NA | 0.1 |
* Class: FUNG = fungicide, HERB = herbicide, NEO = neonicotinoid, PS = partially systemic, S = systemic.
LOQ = limit of quantification (µg/L).
Mean and SEM for detections> LOQ.
Pesticide | Class |
Detection | Samples (N) | Proportion of | Concentrations (µg/L) | LOQ |
|||
positives (%) | Min | Max | Mean |
SEM |
|||||
Clothianidin | NEO, S | 34 | 34 | 100.0 | 0.0170 | 2.3000 | 0.523 | 0.567 | 0.001 |
Thiamethoxam | NEO, S | 34 | 34 | 100.0 | 0.004 | 2.8 | 0.585 | 0.632 | 0.0001 |
Azoxystrobin | FUNG, S | 21 | 34 | 61.8 | 0.001 | 2.1 | 0.191 | 0.587 | 0.001 |
Imidacloprid | NEO, S | 3 | 34 | 8.8 | 0.001 | 0.007 | 0.004 | 0.003 | 0.001 |
Imidacloprid urea | NEO, S | 3 | 34 | 8.8 | 0.005 | 0.005 | 0.005 | 0 | 0.0009 |
* Class: FUNG = fungicide, HERB = herbicide, NEO = neonicotinoid, PS = partially systemic, S = systemic.
LOQ = limit of quantification (µg/L).
Mean and SEM for detections> LOQ.
Comparison of mean concentrations of clothianidin and thiamethoxam potentially ingested per honey bee with their respective oral LD50 values revealed a mean acute risk quotient (RQ) below 0.1 for samples collected during corn planting (
Neonicotinoid | AOT LD50 (ng/bee) |
Planting | Samples (N) | Concentrations in water (µg/L) | Body burden in bees (ng/bee) |
RQ |
|||
Mean |
Max | Mean |
Max | Mean |
Max | ||||
Clothianidin | 3.35 | During | 25 | 4.6 | 55.7 | 0. 21 | 2.62 | 0.06 | 0.78 |
After | 34 | 0.5 | 2.3 | 0. 02 | 0. 11 | 0.01 | 0.03 | ||
Thiamethoxam | 4.4 | During | 25 | 7.7 | 63.4 | 0. 36 | 2.98 | 0.08 | 0.68 |
After | 34 | 0.6 | 2.8 | 0. 03 | 0. 13 | 0.01 | 0.03 |
* Acute oral toxicity (AOT) values at 24 hours
Conversions are based on the drinking water intake rate of 0.047 ml (EFED & PMRA 2012).
RQ = Risk Quotient.
Mean for detections> LOQ.
Neonicotinoid seed dressing is used extensively in agriculture to protect a wide variety of crops from pests. As these insecticides are highly toxic to honey bees, it is essential to identify and quantify every potential route of exposure. Field observations of honey bees drinking from puddles of rainwater raised concerns about their potential exposure to these systemic compounds.
The results presented here more clearly define a previously uninvestigated route by which honey bees are exposed in corn-dominated environment, not only to neonicotinoid insecticides, but also to a cocktail of herbicides and fungicides (
While the average acute risk associated with consumption of puddle water alone was found to be relatively low for pollinators (
Risk assessment for contact with and dietary exposure to pesticides is a process thoroughly described for honey bees
Water collection depends entirely upon the colony's demand, since water is not stored inside the hive
To our knowledge, this is the first scientific record of neonicotinoid residues in in-field puddles of water in relation with neonicotinoid seed dressing in corn cropping system. Although concentrations of these systemic insecticides in water samples were not found to be above lethal doses, repeated exposure through consumption of puddle water alone can result in various sublethal effects at the individual- and colony-level. Moreover, due to the abundance of water puddles in agriculture-intensive areas and their particularly attractive features for honey bees, they are highly likely to be one of the main, and at times exclusive, supply of water and thus an important source of pesticide exposure Finally, we believe that the risk of exposure to neonicotinoid-contaminated water reported here is an underestimation. Additional, comprehensive research is needed to therefore better assess risk associated with water use for honey bees. Our findings provide further evidence of the widespread environmental contamination with neonicotinoids and highlight another potential route of exposure for honey bees and other pollinators.
(DOCX)
(CSV)
The authors would like to thank Luc Gagnon of the Ministère de l'Agriculture, des Pêcherie et de l'Alimentation (MAPAQ) and Christian Deblois of the Centre d'expertise en analyse environnementale du Québec (CEAEQ) for their assistance with chemical analyses. Special thanks to Étienne Nadeau for his assistance with sample collection during the second year of the study. We are also thankful to the CÉROM for providing supplies and technical assistance. We thank Connie Hart and Wayne Hou of the Pesticide Management and Regulatory Agency (PMRA) for their assistance with pesticide exposure estimates. We are also thankful to anonymous reviewers whose comments and valuable insights helped improve the manuscript. Finally, O. S.-R. would like to thank the Quebec's Center for Biodiversity Science and NSERC-CANPOLIN for various scholarships.