In Hungary, colony numbers fell by 24% in the following spring. In the UK, honeybee colony survival was generally very low, but lowest where bees fed on clothianidin treated oilseed rape in the previous year.
No harmful effects on overwintering honeybees were found in Germany.
The experiment – undertaken in the UK, Germany and Hungary by researchers from the Centre for Ecology & Hydrology (CEH) and published in the peer-review journal Science – exposed three bee species to winter oilseed rape crops treated with seed coatings containing neonicotinoid clothianidin, from Bayer CropScience, or Syngenta’s thiamethoxam.
HTA horticulture head Raoul Curtis-Machin said the results are not relevant to ornamentals production in the UK: "Clothianidin has never been approved for use on plants for gardens and landscapes, and whilst thiamethoxam has a current EAMU approval for use on non-flowering crops that are grown under protection and had approval in the past for ornamentals use, this is no longer the case on any flowering crop and has had no approval since 2014."
Neonicotinoid seed coatings are designed to kill pests such as the cabbage stem flea beetle, but were effectively banned in the EU in 2013 because of concerns regarding their impact on bee health. Some scientists have argued their effects persist in the environment.
Crop Protection Association chief executive officer Sarah Mukherjee said: "The fact that bees fared better in Germany than in the UK and Hungary suggests that neonicotinoids are but one of many factors which impact on bee health. There have been a multitude of changes in their environment since before the invention of neonicotinoid pesticides; habitat loss, parasitic mites, agricultural and beekeeping practices and climate change. Blaming neonicotinoids without addressing these other factors will do very little to improve the viability of bee populations.
"As the researchers themselves acknowledge, neonicotinoids have many positive attributes – they are targeted, applied at low dosage rates, reduce the need for spraying and are useful in controlling pests where resistance to other pesticides has been found. The report’s authors agree that neonicotinoids are an important part of the farmer’s toolbox, helping them protect their crops and ensuring a supply of healthy, safe and affordable food."
NFU vice president Guy Smith said: "Both bees and farmers play a crucial role in producing food which is safe, affordable and high quality. Farmers rely on bees to pollinate crops and have planted around 10,000 football pitches of flower habitat across the country to support a healthy bee population and give them a good home – all because they recognise the key role they play. The oilseed rape crop itself is a great early source of food for them too.
"We strongly believe that policy decisions – such as restricting the use of neonicotinoids – must be based on sound science which gives strong evidence. And while this CEH study provides more useful information, we still don’t have that definitive evidence for the impact of neonicotinoids, and yet farmers and food production are hindered by restrictions on these products.
"Carrying out the task of producing food for the nation and the world depends very much on how healthy a crop is and how well we can protect it from pests. Insecticides – neonicotinoids in this case - are an important part of protecting not just oilseed rape but also crops such as cereals, sugar beet and vegetables. Without the tools to protect our crops, such as neonicotinoids which remain widely used all over the world, we risk simply importing food from other countries whose production standards we can’t control."
Dr Chris Hartfield, NFU lead on bee health, said: "The NFU has called for large scale field trials like these to be done to try and resolve the question of whether harmful effects seen in artificial-feeding studies actually occur in real fields, to the extent where neonicotinoids could be causing widespread decreases in bee populations.
"The largest result from this study is that for over 75% of the parts of bee life cycle measured, the neonicotinoid seed treatments had no significant effect at all. Where significant effects did occur, there was an inconsistent range of negative and positive effects of neonicotinoid use on bee health. The study found negative effects of neonicotinoids on bees in Hungary and the UK, and positive effects on bees in Germany and the UK. If neonicotinoids were as bad as some make out, their impact would not respect geographical boundaries. There would be clear and similar impacts in every country and every field where neonicotinoids were used. This study clearly shows this is not the case.
"The report makes clear these effects are not consistent across countries and suggests this is because there are many interacting environmental factors that impact on bees, not just neonicotinoids. I think this also indicates that effects of neonicotinoids, if they occur at all, are small and can be overridden by other factors, such as plentiful food resources and good colony health.
"This study does not show that neonicotinoids are causing widespread declines in bee populations – particularly given that it shows positive as well as negative effects of neonicotinoids on bees, and in the majority of cases there is no effect at all. And let’s not forget the wider backdrop that it is still the case that the evidence shows the most significant declines in bee biodiversity in the last century actually predate the use of neonicotinoids by decades. More recently – during the time neonicotinoids have been used – there has been a slowdown in the rate of decline in bumblebee biodiversity in the UK, and also some increases in solitary bees in Great Britain."
In the study, lower reproductive success – reflected in queen number (bumblebees) and egg production (red mason bee) – was linked with increasing levels of neonicotinoid residues in the nests of wild bee species buff-tailed bumblebee (Bombus terrestris) and the Red Mason Bee (Osmia bicornis) across all three countries.
According to the CEH lead author Dr Ben Woodcock: "The neonicotinoids investigated caused a reduced capacity for all three bee species to establish new populations in the following year, at least in the UK and Hungary."
He suggests the differing impacts on honeybees between countries may be associated with interacting factors including the availability of alternative flowering resources for bees to feed on in the farmed landscape as well as general colony health, with Hungarian and UK honeybees tending to be more diseased.
In contrast, the hives in Germany happened to be larger, showed little evidence of disease and had access to a wider range of wild flowers to feed on. Dr Woodcock suggests that this may explain why in this country alone there was no evidence of a negative effect of neonicotinoids on honeybees.
The study spanning 2,000 hectares, equivalent to 3,000 full scale football pitches, took account of bee disease and surrounding landscape quality in addition to colony growth rate, worker mortality and overwinter survival.
Woodcock said: "Neonicotinoid seed dressings do have positive attributes: they target insects that damage the plant, can be applied to the seed at low dosage rates but protect the whole plant and reduce the need for broad spectrum insecticide sprays. Their use as an alternative chemical control option is also useful in controlling pests where insecticide resistance to other pesticides is already found, so play an important role to play in food production."
He added: "There may be opportunities to mitigate negative impacts of neonicotinoid exposure on bees through improved honeybee husbandry or availability of flowering plants for bees to feed on across non-cropped areas of the farmed landscape. Both these issues require further research.
"The negative effects of neonicotinoids on wild bees may also be the result of diverse mechanisms of exposure that include persistent residues of neonicotinoids in arable systems due to their widespread and often very frequent use."
Co-author Professor Richard Pywell, Science Area Lead, Sustainable Land Management at the Centre of Ecology & Hydrology, said, "Neonicotinoids remain a highly contentious issue with previous research on both honeybees and wild bees inconclusive.
"This latest field study was designed, as far as possible, to reflect the real world due to its size and scope. We therefore believe it goes a considerable way to explaining the inconsistencies in the results of past research, as we were better able to account for natural variation in factors like exposure to the pesticide, bee food resources and bee health for different bee species.
"Our findings also raise important questions about the basis for regulatory testing of future pesticides."
Bayer CropScience and Syngenta funded the research assessing the impact of neonicotinoids on honeybees. The Natural Environment Research Council funded the analysis of the impact on the wild bees. The experiment, including design, monitoring and analysis, were scrutinised by an independent scientific advisory committee chaired by Professor Bill Sutherland of Cambridge University.
Prof Charles Godfray, Hope Professor of Entomology at the University of Oxford, said:
"The two papers published in Science this week materially add to the evidence base we have to make decisions about the use of neonicotinoid insecticides in agriculture. They both demonstrate that these chemicals can harm pollinators, though their effects are variable across species and location. Interestingly, they suggest that pollinator intake of insecticide through non-target plants (weeds etc., growing in the vicinity of crops), and the persistence of neonicotinoids in the environment, may be greater than previously thought."
Dr Giles Budge, Science Lead in Crop and Bee Health at Fera, said:
"This study is a laudable attempt to assess the impact on pollinators of landscape neonicotinoid usage in real farmed environments across three countries. Conducting a study of this magnitude is technically very challenging and there will always be subtle differences between geographical regions. Overall the results suggest that the use of neonicotinoids can produce both negative and positive impacts on wild and managed pollinators depending on the geographical context of usage.
"In Germany, neonicotinoid usage was positively associated with honey bee health, but in Hungary the association was negative. Interestingly, in the United Kingdom the results were variable with both positive and negative impacts recorded - although the report of high overwinter colony mortality in this study (Control 58%; Clothianidin 79% and Thiamethoxam 67%) is a concern when the annual overwinter losses in 2014-15 were far lower at 14.5%. The reasons for this are not explained but it is worth noting these colonies were carrying high Varroa mite loads overwinter which is a known driver of colony losses.
"Positive and negative impacts on bumblebees were also recorded, with the presence of treated fields being a poorer predictor of pollinator health than when chemical exposure in the nest was quantified, suggesting the context of local non-crop forage is important. Overall this study provides a useful addition to the literature and highlights the need for landscape-scale experiments assessing the drivers of poor pollinator health to consider multiple stressors, such as forage availability, climate and parasite pressure."
Prof David Goulson, Professor of Biology at the University of Sussex, said:
"This is by far the largest field trial ever conducted on the impacts of neonicotinoids on bees, including honeybees, bumblebees and solitary bees, and conducted simultaneously across the UK, Germany and Hungary. Exposure of bees was entirely field-realistic; indeed, farmers simply followed normal farming practice.
"The findings are in agreement with a number of earlier studies; field exposure to neonicotinoids has clear negative impacts on bumblebees and solitary bees. Effects on honeybees were also predominantly negative but more variable. Interestingly, analysis of residues of neonicotinoids in bee nests suggests that much bee exposure was not from the treated crop adjacent to the colony but was coming from other sources in the landscape, suggesting widespread contamination of the environment. For example, the neonicotinoid imidacloprid was frequently detected in bee nests but was not used on the farms in the year of the study. This is in accordance with previous studies showing that neonicotinoids are persistent in soils and frequently contaminate wildflowers.
"In the light of this new study, continuing to claim that use of neonicotinoids in farming does not harm bees is no longer a tenable position".
Prof Lin Field, Head of Department of Biological Chemistry and Crop Protection at Rothamsted Research, said:
"The paper reports a set of experiments looking at the effect of two neonicotinoid OSR seed treatments (thiamethoxam and clothianidin) on three bee species (honey bee and two wild bees B. terrestris and O. bicornis) in three countries (Germany, Hungary and the UK).
"I am not an expert in the setting up of field trials and their statistical analysis but I am happy to assume that these are sound. Without seeing the Supplementary material (which are presumably very large data sets) it means we take the summary of results in Figure 2 at face value.
"The results are very mixed. There are 42 comparisons of parameters measured with and without the seed treatments and of these, 33 are not significant and therefore we can only conclude that the treatments had no effect. For those that are significant there are two positive effects in the results from Germany, three negative effects in Hungary (assuming that both CTD and TMX are significant) and three negative and one positive effect in the UK. None of these is consistently positive or negative for any one parameter.
"Although there is no significant effect of either compound on queen production in B. terrestris, in Figure 2, Figure 3 appears to show a negative correlation between queen production, in both wild bee species, and the total amount of neonicotinoid found in nests. One of the worrying things about the data in Figure 1A is that if each country is looked at separately, Hungary shows a wide range of values for queen numbers with very little variation in nest residues, whereas the UK shows a wide range of residue levels but little variation in queen number. It is hard to interpret this.
"So overall most of the parameters tested show no significant differences and for those that do there are sometimes conflicting results, meaning that it is hard to draw any conclusions. The authors do make some comments on what the negative effects might mean for bee populations more widely but do not comment on the positive effects, which are hard to explain. My feeling is that no firm conclusions can be drawn from these results. This is perhaps not surprising given that there will be many complicated interactions between different neonicotinoids, different bee species and different field conditions."
Dr Christopher Connolly, Reader in Neurobiology at the University of Dundee, said:
"This industry-funded study supports the large body of evidence against the neonicotinoids by demonstrating a negative impact on bees that is most striking for bumblebees and solitary bees.
"The major impact of the neonicotinoids on bees is known to be by the disruption of brain cell function, weakening the bee’s ability to learn and remember, and so reduce their foraging efficiency and ultimately, colony performance.
"The important point about this mechanism of action is that it should be possible to mitigate against these effects by the provision of alternative, non-contaminated, food sources such as bee-friendly gardens and natural ‘wilderness’ habitats
"Providing more wildflowers has two direct benefits to bees, the reduction in exposure to crop chemicals and to make the finding of forage much easier. Therefore, the impact of neonicotinoids will depend heavily on the local habitat, making it unsurprising that a large variability in toxicity is found in this study.
"In addition to exposure to the experimental neonicotinoids, each treatment group was also exposed to other pesticides (fungicides and insecticides) that differed between groups. So, like the real world, an understanding of the cause and effect is confounded by the use of other agrochemicals (see also companion article in Science by Tsvetkov). Furthermore, the real-world problem of the ubiquitous presence of pesticides was demonstrated by their detection of another neonicotinoid (imidacloprid) that was not part of the study. Indeed, this natural neonicotinoid contaminant may have been responsible for driving the effects on bumblebees and solitary bees.
"As imidacloprid is subject to the EU moratorium (since 2013) preventing the use of neonicotinoids on bee-visited crops, this study allows one further important conclusion – this partial EU ban on the use of neonicotinoids does not prevent their exposure to bees."
Norman Carreck, Laboratory of Apiculture and Social Insects at the University of Sussex, said:
"This is the largest field experiment carried out to date studying the effects of neonicotinoid seed treatments on bees. As with most previous field studies, the results clearly demonstrate the practical difficulties of performing such experiments, and whilst adding to our knowledge, the study throws up more questions than it answers.
"The results are inconsistent, and very few differences are statistically significant. For honey bees, both negative (in Hungary) and positive (in Germany) effects were observed. It is unfortunate that at the UK sites, small nucleus honey bee colonies were used rather than full sized colonies, and the majority of these died out during the course of the experiment. Only 23 out of a total of 72 hives survived, and the control colonies which had no access to the neonicotinoid treated crops suffered 58% mortality, making it difficult to draw any reliable conclusions. For the wild bumble and solitary bees, few significant effects could be attributed to the crop treatment, but some adverse effects were correlated with residues, presumably from previous cropping.
"A consequence of the EU moratorium since 2014 has been that farmers in southern and eastern England have suffered increased pest problems, especially of the cabbage stem flea beetle, and as a result have greatly increased the use of older chemical compounds, most notably synthetic pyrethroids, with unknown consequences on bees and other beneficial insects. Further studies are needed to assess the effects of the moratorium on bee populations."
Dr Philip Donkersley, Lancaster University, said:
"This study is particularly novel, in that it has concordantly assessed the non-target effects of field-level exposure to two key neonicotinoid insecticides in a field setting across three distinct EU countries. This study is also important because it addresses not only honey bees, but also the less well studied wild (bumble and solitary) bees.
"The conclusion of the study is that pollinating insects have different responses to neonicotinoid exposure in different EU countries. This clearly indicates the importance of context-specific conditions that complicate any conclusions made on experiments of this scale.
"The authors state that the effects of neonicotinoids they observed are likely ‘a product of interacting factors’ such as ‘differences in the use of oilseed rape crop as a forage resource for bees’ and ‘incidence of disease within hives’. One issue I have with this manuscript is that the authors do not attempt to account for these factors in their analysis, and thus cannot provide a stronger conclusion on these effects.
"As different countries show different effects, one could suggest that an EU-wide moratorium on neonicotinoids may not be the best approach. Although the multi-national cooperation was the most effective response to preliminary reports on bee losses, evidence is mounting that country-specific legislation may be more effective at protecting pollinators."
Emeritus Prof Rob Smith, University of Huddersfield, said:
"This is a well-designed study carried out an appropriate scale to detect field effects of neonicotinoid seed treatments. It makes an important contribution to the debate about effects of particular insecticidal treatments on honeybees and wild bees.
"The interpretation is not, however, clear-cut because treatment effects do not just differ between the three countries but, in the case of honeybees, produce within-season effects that appear to be negative in two countries and positive in one. Between-year effects on measures of hive fitness appear to be highly variable; this study, while large, does not seem to have the power to detect between-year effects (only one out of 24 control vs. treatment comparisons is statistically significant).
"Turning to the two wild bees, neither species was directly affected by the experimental seed treatments. Queen production in one and reproductive cell production in the other were negatively correlated with a measure of neonicotinoid residues in the nest, but their measure of nest residues does not relate closely to the experimental treatments because it introduces a third neonicotinoid not applied in the study. No data are given on residues of any other compounds (which would have been used during the EU moratorium on neonicotinoids) and these may be important confounders.
"In summary, these results are important in showing that there are detectable effects of neonicotinoid treatments on honeybees in the real world. But these effects are not consistent between countries and, in this study, appear to be both positive and negative in different places.
Prof Nigel Raine, Rebanks Family Chair in Pollinator Conservation at the University of Guelph, said:
"This long awaited paper reports important results from the largest scale field study to date to assess the impacts of neonicotinoid exposure on insect pollinators. The authors compared the performance of three bee species (honeybees, bumblebees and a cavity nesting solitary bee species) across 33 sites in three European countries (UK, Germany and Hungary) when exposed to oilseed rape fields treated with either the neonicotinoid insecticides clothianidin or thiamethoxam, or fields where no neonicotinoid was applied."
"Whilst results from this large scale study report varying impacts of neonicotinoid exposure both among bee species tested and countries assessed, the overall picture points towards appreciable negative impacts on these important pollinators across the time course of this study. Perhaps most novel in this study (and that by Tsvetkov et al. published in the same issue of Science) are the measurable and significant impacts on honeybee individuals and colonies. Previous field studies have not detected such impacts on honeybees, and the ambitious scale of this study could be a significant factor in why measurable impacts were found on this occasion."
"It is concerning that bumblebee colonies produce fewer queens, and solitary bees (Osmia bicornis) produce fewer offspring, where higher levels of exposure to neonicotinoids were found. These bees represent the basis for the next generation of these species in the following year, and fewer of these important individuals could have significant impacts on population size and persistence."
"The authors detected a neonicotinoid, imidacloprid, in their samples that was not used as a seed treatment in this experiment. This suggests that residues from previous agricultural applications could still be affecting bees in the field even several years after the EU moratorium on these active ingredients (imidacloprid, clothianidin and thiamethoxam) came into effect."
"Results from this study, and Tsvetkov et al published in the same issue of Science, provide additional support for restrictions on the use of neonicotinoids based on concerns about impacts on insect pollinator health. Such regulations must balance the benefits of using insecticides to control damaging crop pests appropriately against the unintended costs of harming beneficial insects exposed to these chemicals in agricultural landscapes. Pollinators are responsible for one in three mouthfuls of food we eat, so safeguarding their health is something we should all care deeply about."
Dr Peter Campbell, Senior Environmental Risk Assessor at Syngenta (which manufactures and sells products containing the neonic thiamethoxam), said:
"The one line simplistic summary conclusion published does not reflect the data presented in this paper, which clearly show no consistent effect of neonics across the 3 countries studied, with both positive beneficial effects and negative effects reported for both honeybees and bumble bees, in Germany and Hungary/UK respectively.
Potential explanation for the inconsistent country specific effects reported for Honeybees
"This CEH paper does not present the full set of data analysis conducted by CEH and reported to Syngenta for honeybees. For example the pre-winter data analysis carried out by CEH which showed that any effects reported during the flowering period had disappeared (i.e. recovery), were not included in the paper. There were in fact 258 separate honeybee statistical data analyses reported to Syngenta by CEH. Out of these analyses, 238 resulted in no effect, 7 resulted in beneficial effects, 4 with insufficient data and only and 9 resulting in negative effects. The rules for statistical significance allow for a 5% probability of generating random effects. Therefore based on this internationally accepted statistical benchmark and the 258 analysis CEH carried out, we could expect 13 random results. Therefore the –ve and +ve results reported by CEH could easily be random i.e. not real, and a conclusion of no effect of the neonics reached. It should also be noted that the pollen and nectar residue analysis reported by CEH in this paper indicated that circa 95% of the time no neonic residues were measured, even in samples taken directly from the treated crop. Therefore bees in these trials were hardly ever exposed to any neonic residues.
"The correlation graphs reported for both wild bee species must be treated with caution. Correlation is not causation. Particularly as both graphs show a high degree of natural background variability represented at the lowest residue levels. There are also too few samples at the high residue end of the curve which actually heavily influence these correlations, to show if this variability would be similar or significantly different from the variability reported for low residues. This high level of natural variability in Bumble Bee Queen production and Osmia cell production, even for bees never exposed to neonics, is well known but not fully understood. Potential causes of this variability in this study are:-
- use of different sub-species of Bumble bees in UK compared to Germany and Hungary. This could explain the clear country specific differences observed for bumble bees queen production in these graphs. I.e. if the UK data was removed from the bumble bee analyses the correlation would disappear.
- habitat specific landscape variations between and even within countries
- the female emergence success of the solitary bee cocoons introduced into nests at beginning of the study, before exposure to neonics, was not measured. This is emergence is known to be variable and could easily effect the numbers of cells produced
- Only 2 out of 6 Osmia nests placed in each field were included in this CEH data analysis. From previous studies and our own work we know that Osmia is gregarious, i.e. we have seen significant variability in individual nest occupancy in previous studies that could have happened here across all 6 nests at each site and thus affected the analysis.
"Regardless of the important factors listed above which could have confounded the CEH analysis, we also need to consider what the implications of these reported analysis are at face value. Interestingly both positive (beneficial) and negative country specific effects are reported for honeybees and bumble bees. This demonstrates that neonics can be used safely or even with benefit to bees under certain circumstances e.g. such as reported in Germany. We need to understand what factors are driving the beneficial effects of neonics reported in this study so as to identify and promote these elsewhere. CEH suggest the interacting factors are likely to be landscape/habitat, which may affect availability of alternative wild forage for bees, and for honey bees, good beekeeping husbandry.
"Through our Operation Pollinator Program and Good Growth Plan we are already well placed to research further, and help promote, the uptake of bee friendly landscape management practices.
With regards to the correlational effects reported for wild bees with total neonic residues, it should be noted that CEH acknowledge that the ecological significance of these correlations remains uncertain. However, interestingly once again we see a country specific difference in these correlations, which needs to be fully understood. For example, neonic use pattern differences and landscape and habitat differences need to be investigated and more fully understood to help identify the conditions which will promote sustainable use of these products."