A Buzz in the Meadow

by Dave Goulson

  Science correctly anticipated that these papers, published together, would cause quite a stir. They decided to organise a press conference. The external-relations staff at the University of Stirling were quite taken aback by the idea and seemed less than enthusiastic about hosting it, so the press conference was arranged to take place in Paris instead. Press releases went out at midnight on Sunday 25 March 2012, with a strict embargo on publication of media stories until the evening of Thursday 29 – it was all terribly exciting, for someone who had not been involved in such things before. The press conference was scheduled for Thursday morning. I expected all hell to break loose on Monday, but not much happened. It took a while for the media to react, perhaps because they knew they couldn’t publish anything before the Thursday night, which basically meant that the story would run in the newspapers on Friday.

  After the press release went out, but before the actual papers were available to read, Defra (the Department for Environment, Food & Rural Affairs) declared that all pesticides licensed for use in the UK were perfectly safe, if used properly, and that the two new studies did not change this. This struck me as a remarkably odd position for a government department to take, particularly since it had not at the time actually been able to read the new studies. Its position seemed to reflect a belief that the systems it had in place for evaluating the safety of pesticides were infallible, and therefore that no new evidence could possibly come to light which could reveal that they had made a mistake. One might expect, and even understand, such a stance from the agrochemical companies, but from a government department – paid for by taxpayers and supposedly working on our behalf – it was mystifying.

  As the week went on my phone became steadily busier, with reporters from various newspapers ringing for additional detail and quotes. Penelope and I flew to Paris on the Wednesday evening. We’d had to book a hotel way out in the suburbs as there happened to be an art festival going on in the city centre and everywhere was booked up. On the Thursday we caught the metro into the centre of Paris. The press conference was in a beautiful old building just off the Champs-Elysées, in a room that seemed more suited to a masked ball than a press conference about pesticides. Penelope and I sat nervously next to the French scientists, Mickaël Henry and Axel Decourtye, facing an audience of journalists flanked by television cameras and bright spotlights, while French aristocrats gazed imperiously down from huge gilt-framed oil paintings. We both felt like fish out of water, and were worried that we might face aggressive questions from representatives of the agrochemical industry. Penelope was heavily pregnant at the time, and we joked that she should feign contractions if things weren’t going well. In fact the conference was reasonably uneventful, although some of the questions were a little eccentric. One Mexican journalist suggested that we should develop a vaccine against neonic poisoning. Afterwards we did interviews for a couple of television stations and then escaped to a park by the Seine, where I spent the whole afternoon on my mobile fielding questions from journalists.

  The next day, back in the UK, the newspapers were full of the story. The Independent, always a staunch supporter of bees, had put the story on the front page, while every other broadsheet and most of the tabloids had covered it in one way or another. The following day Defra’s Chief Scienctific Advisor, an eminent scientist named Robert Watson, announced that Defra would review the evidence with regard to neonics. We were delighted.

  In the following weeks we heard that Defra had commissioned Fera (the Food & Environment Research Agency) to repeat our study, but that they were going to expose their bumblebee nests by placing them next to real fields of rape. I couldn’t see how they could find control (untreated) fields, particularly since they seemed to be starting the study immediately, but I assumed they had rather greater resources at their disposal than we had. The French Agriculture Minister announced an immediate ban of thiamethoxam (the chemical studied by the Avignon research team) on oilseed rape. The European Commission asked EFSA (the European Food Safety Authority) to conduct a thorough review of the safety of neonics. In addition EFSA launched a review of the safety tests used to evaluate new agrochemicals, particularly with a view to devising ways to detect sub-lethal effects on bees. The Environmental Audit Committee in Westminster launched an inquiry into the safety of neonics. It seemed that the science was being taken seriously by policymakers, and that wider restrictions on the use of neonics would be forthcoming.

  In the meantime, research on neonics and bees continued. A paper from James Cresswell’s research group at Exeter University showed that doses of imidacloprid as low as one part per billion were enough to reduce egg-laying in bumblebees by one-third. Nigel Raine’s group at Royal Holloway published a study showing that one of the main effects of imidacloprid on bumblebees was that the workers collected far less pollen. These studies dovetailed neatly with our own; the slow growth and reduced queen production of our nests could readily be explained if the workers brought back less food and the queen laid fewer eggs. By the autumn of 2012 the evidence that neonics were likely to be having a major impact on wild bumblebee populations was coherent and convincing.

  Sometime in October I received an anonymous email from the United States. It suggested that I should read a document that was attached. The document was entitled ‘Draft Assessment Report: Initial risk assessment provided by the rapporteur Member State Germany for the existing active substance Imidacloprid, Volume 3, Annex B, February 2006’. If the title sounds dull, I can assure you that it was nothing compared to the document itself, and I was sorely tempted just to delete it, but my interest was sufficiently piqued that I eventually set about trawling through it. It comprised pages 572–790 of a much larger report, and it summarised many dozens of scientific investigations carried out by the manufacturers of imidacloprid, most of them relating to its chemical structure and the chemistry and speed of its degradation in soil. It was mind-numbingly tedious, but eventually a graph, some seventy pages in, caught my eye. It described the results of a study conducted by the pharmaceutical company Bayer on the levels of their product, imidacloprid, in soils over a six-year period in the early 1990s. They had simply sown winter wheat treated with imidacloprid for six years in a row at two sites in East Anglia, and then measured the level in the soil the day before the following application. The data were absolutely clear – the levels simply went up and up over time, up to 60 parts per billion, far higher than the concentrations we had been using with our bees. It was abundantly clear that imidacloprid has the potential to accumulate in soil if used regularly. Yet the text describing this study concluded that it demonstrated ‘no potential for accumulation in soil’. What on earth was going on? I sent the document to the Environmental Audit Committee and they quickly got back to me with the opinion that it must be a fake, perhaps produced by a rabid anti-pesticide campaigner. However, further investigation showed it was genuine – and that anyone could download the original from an EC website, if they had the patience and knowledge to find it, tucked away amongst thousands of other lengthy and tedious documents.

  This got me thinking. Up until this point I had been focused on bees, but had I been missing the bigger picture? I began to dig more deeply into the literature about the environmental fate of these compounds. The first thing I discovered I found quite astonishing. A study by Bayer scientists had quantified exactly how much of the active ingredient that stuck on the outside of crop seeds was taken up by the crop. The answer was: not very much – usually only about 2 per cent, and sometimes up to 20 per cent. Other studies carried out in Italy showed that about 1 per cent of the seed coating routinely blew away as toxic dust, even when the pesticide was firmly stuck to the seeds; not very much, but enough to kill immediately any honeybees flying nearby. The vast bulk of the chemicals, between 80 and 98 per cent, were ending up in the soil. The agrochemical industry had always claimed that pesticides applied as seed dressings provide much better targeting of the crop compared to those applied as sprays, but this does
not appear to be true. With spray applications one can commonly get 30–50 per cent of the active ingredient on to the crop; neonic seed dressings appear to be far less efficient in this respect.

  Disturbingly, studies of the persistence of neonics once in soil suggested that they could last for years. This is usually measured as a half-life – the time it takes for half of the chemical to break down – and most published estimates of the half-life put it at anywhere between 200 and 6,000 days, depending on soil type and conditions. This made sense of the East Anglian study; if quite a bit of the chemical is left in the soil after one year, then adding more every year is going to cause levels to rise over time. What does this do to invertebrates that live in soil?

  To make matters worse, neonics are soluble in water – they have to be, to act systemically in plants. This would lead one to expect them to leach from soils into surrounding streams and ponds. The evidence seemed to suggest that there would be a flush of neonics washing out of soils if it rained soon after the seeds were sown (highly likely with autumn-sown crops such as winter wheat), with the remainder of the neonics binding to organic matter in the soil and then being likely to remain there for months or years. If neonics end up in clear water they are quickly broken down by sunlight, but if they settle in pond or stream sediments they can last for years. Random sampling of streams in California commonly found concentrations exceeding one part per billion – higher than the LC50 for some aquatic insects such as mayflies. Studies from the Netherlands suggest that heavy use of neonics on the bulb fields can lead to concentrations exceeding 200 parts per billion in nearby waterways, presumably sufficient to kill all insect life. I could find no similar studies from the UK of the levels in water.

  So, we can expect neonics to have been gently accumulating in arable soils throughout the world for the last twenty years. It seems likely that they will be taken up by hedgerow and field-margin plants that have their roots in these soils, just as they are by the crop, meaning that they might be consumed by any farmland herbivore – the caterpillar of a butterfly, for example. Do the nettle patches in field margins all contain neonics, so that the peacocks and small tortoiseshell caterpillars that eat them all become poisoned? Are all the moth caterpillars feeding on the hedgerow trees slowly accumulating toxins? Even if not sufficient to kill them, does this exposure interfere with their behaviour as adults, making them less able to find mates or less adept at identifying the correct places to lay their eggs? Are the field margins and flower strips paid for by agri-environment schemes contaminated with neonic dust from seed-drilling and uptake from the accumulations in soil? Neonics are also likely to be pulsing into streams in autumn when most crops are sown, and building up in aquatic sediments. What harm does this do to aquatic insects and the fish that depend on them for food? Finally, do birds and rodents eat treated crop seeds? Even though neonics are less toxic to vertebrates than to insects, a single maize seed is coated with enough neonic to kill a songbird, and seeds are commonly spilled onto the ground during sowing operations. It may be that we are poisoning the environment on a monstrous scale. Agri-environment schemes and other conservation projects are doomed to fail, if this is so. Could all of this explain why farmland wildlife is declining, despite our best efforts to look after it? I don’t know, but it seems entirely plausible, and it is surely high time we found out. In the meantime it seemed to me that it might be wise to stop using neonics.

  In December 2012 EFSA announced the results of its review of the safety of neonics. It seemed to agree with me; it highlighted that many of the environmental risks posed by these chemicals were not properly understood, and concluded that neonics posed an unacceptable risk when used on any crop visited by honeybees, or for crops sown when bees are active (due to the toxic dust created). A few days later the EC proposed a two-year moratorium on all such use of neonics, which was to be put to a vote of member states.

  Of course the agrochemical industry didn’t take this lying down. Global neonic sales are thought to be worth roughly $3.5 billion, so the people who make them have an awful lot to lose if they are banned. Our research, and that of the Avignon group, came under attack. Industry claimed that we had used unrealistically high doses, and that our work was lab-based and thus not representative of the real world. They claimed that EFSA’s six-month review was shoddy and rushed. They produced glossy documents claiming that, if a ban on neonics was introduced, the EU economy would suffer to the tune of seventeen billion euros, and that 50,000 jobs would be lost, although it was unclear what evidence underlay these statistics. The document seemed to be designed to scare politicians into voting against the moratorium.

  At around this time, in January 2013, I was asked to go to the Central Science Lab in York for a meeting to discuss the findings of the Fera study. This was the study that had attempted to copy our research on bumblebees and neonics, but with the exposure part of the study taking place in the field, so that the entire experiment was as realistic as possible. Fera scientists had placed buff-tailed bumblebee nests next to one of three fields, one treated with clothianidin, one with imidacloprid and a control field that was untreated with pesticides. The intention was to compare how the nests performed; if neonics are harmful to bumblebee nests, then the nests next to the treated fields ought to do worse than those next to the control field. This is not a great experimental design; one really needs several fields in each treatment, since every location will be different in numerous other ways that might confound any effects of the treatment itself. However, Fera could only find one untreated field in the whole of England, and they had had to drive 160 kilometres to get to it. As it turned out, this was the least of their problems. The bigger issue was that bees can fly, a fact of which they ought to have been aware. Buff-tails happily fly a couple of kilometres to find food, and they are suspected of avoiding foraging close to their nests, perhaps because it might attract predators. When the Fera scientists tested the food stores in the nests for pesticides, it turned out that the ‘control’ nests contained a selection of neonics. The bees had clearly found a crop treated with thiamethoxam, and another treated with clothianidin (or perhaps wild flowers contaminated with both) and had been feeding on these. The control nests were exposed to just as much pesticide as the ‘treated’ bees. This was a disaster, and it might have been wise to abandon the experiment at this point and start again. All that could really be learned was that bumblebee nests in UK farmland are heavily exposed to a cocktail of chemicals, no matter where you put them. Instead, Fera or their Defra bosses chose to ‘publish’ the study by placing it online, and declared that it demonstrated there was no major effect of pesticides on bumblebees. The normal scientific process is to submit studies to a peer-reviewed journal, where they are scrutinised by anonymous, independent experts. Weak and flawed studies such as this are weeded out, providing a degree of quality control. Simply placing results on the Internet is not the way science proceeds.

  In fact, if one looks closely at the Fera study, it emerges that there is a clear and strong correlation between the levels of pesticide in each nest and how poorly the nest fared, in terms of growth and queen production. Nests with more pesticide grew more slowly and produced fewer queens. Oddly this was not mentioned in the summary of the paper, or when Defra subsequently referred to the study.

  The EC proposal went to a vote on 15 March 2013; of twenty-seven EU states, thirteen voted for the moratorium, nine against and five (including the UK) abstained. The proposal was one vote short of an absolute majority, and so by EC rules the proposal was rejected. Defra stated that there were still too many unknowns for it to formulate an opinion. With impressive chutzpah, the government’s Chief Scientific Advisor, Sir Mark Walport, stated that everyone else had misinterpreted the evidence, and that in the meantime we should invoke the ‘precautionary principle’ by continuing to use neonics. It cited Fera’s study as showing that there seemed to be no major effect of neonics on bumblebees, even though it was well aware that Fera’s s
tudy was a disaster.

  It was hard for those of us involved to understand the stance of countries that voted against the proposal, or those that abstained. Almost everybody agrees that there are huge gaps in our understanding of the risks associated with these chemicals, and that they were not adequately assessed when first proposed for use. EFSA’s team of scientists concluded that these chemicals pose unacceptable, but poorly quantified, risks to bees. If neonics were brought to the market for the first time now, supported by the statement ‘These chemicals seem to pose substantial risks to the environment; we haven’t properly evaluated these, but we’d like to start selling them all over the world anyway’, they would of course be rejected. But because they slipped through the net and are already in use, there seems to be a reluctance to admit a mistake, or to upset the status quo. Somehow the UK government regarded its own stance as an adoption of the precautionary principle, when everyone else’s understanding was that adopting the precautionary principle would lead us to do the opposite – i.e. stop using toxic chemicals until we could be sure they were safe, rather than vice versa.

  Fortunately the EU Health and Consumer Commissioner, Tonio Borg, decided to bring the proposal back for a second vote on 29 April. Once again the UK voted against the ban, but this time Germany switched to a vote in favour, and the bill was passed by fifteen votes to eight. Whatever one’s views on EU membership, this seems to me to be a situation in which common sense was imposed upon an unwilling UK by our more sensible continental neighbours.