David's blog posts for 2017

Are neonicotinoids killing birds?



Philip Lymbery, chief executive of Compassion in World Farming, has just launched his new book “Dead Zone: Where the Wild Things Were.” In it, he suggests that neonicotinoid insecticides may be contributing to bird declines. This claim has come under heavy fire from NFU and the agrochemical industry. In a recent online article NFU’s Dr Chris Hartfield is quoted as saying:

Evidence from real life field situations did not back up the claims, which are based on research by Dave Goulson, Professor of biology at Sussex University. Professor Goulson has published a paper which claimed a grey partridge would be killed by eating five treated seeds, and a sparrow after two. He also said birds which ate insects were ‘declining more rapidly in areas which use neonics’.

            “Dave Goulson’s theories about neonicotinoids poisoning birds are simply that – theories – and are not backed up by evidence from real life”, he added. In the UK, poisoning of all animals is investigated by the Wildlife Incident Investigation Scheme. If seed-eating farmland birds were being poisoned as a result of eating neonicotinoid treated seed, you would rightly expect this scheme to be finding these incidents. There are no incidents of bird poisoning resulting from the use of neonicotinoids over the last ten years. Promoting theories without the evidence to back them up is only going to damage the cause of pollinators and wildlife, and damage the public perception of science in general.

The paper to which Dr Hartfield refers was a review I published in 2013 in Journal of Applied Ecology (Goulson 2013).  In it, I say the following:

Although neonicotinoids do show relatively low toxicity to vertebrates, we might expect seed-eating vertebrates to be exposed to lethal doses if they consume treated seeds spilled during sowing. Typically, maize seeds are each treated with ~1 mg of active ingredient, beet seeds with 0.9 mg, and the much smaller oilseed rape seeds with 0.17mg (Rexrode et al. 2003; Anon 2012; Krupke et al. 2012). A grey partridge, typically weighing approximately 390g, therefore needs to eat ~5 maize seeds, 6 beet seeds or 32 oilseed rape seeds to receive an LD50. A grey partridge typically consumes ~25 g of seeds /day (Liukkonen-Anttila et al. 1999), equivalent to ~600 maize seeds, so clearly there is the potential for birds to swiftly consume a lethal dose. By a similar calculation, 3 maize seeds treated with imidacloprid would deliver more than the LD50 to a mouse. The US Environmental Protection Agency estimate that ~1% of drilled seeds remain accessible to granivorous vertebrates (i.e. they are not buried during drilling), and this does not include spillages which may occur, for example when transporting grain or loading hoppers. With typical sowing rates of ~50,000 seeds/ha for maize and 800,000 seeds/ha for oilseed rape, we might expect sufficient seed to be available on the soil surface to deliver an LD50 to 100 partridge or 167 mice for every hectare sown. 

            Lopez-Antia et al. (2013) fed imidacloprid-dressed wheat seed to red-legged partridge (Alectoris rufa) for 10 days and obtained 58% mortality, with the survivors exhibiting a range of sublethal effects.  This mortality rate, although considerable, is less than we might expect from the calculations above.  Lopez-Antia et al. report anecdotally that partridge did not avoid dressed seed when offered both dressed and undressed, but speculate that treated birds ate less than control birds, and so received a lower dose than expected. This requires further investigation, in this and other species, to determine how much treated seed vertebrates actually consume in the field. De Snoo et al. (1999) describe incidents of poisoning of wild partridge, pigeon and duck by seed dressed with imidacloprid reported by members of the public in France in 1994-5 (a time when neonicotinoid use was very low), but other evidence for effects in the field is lacking and it is unclear whether public reporting is an efficient means of detecting such incidents.

Since 2013, there have been a number of other scientific publications linking neonicotinoids to declines of birds and other vertebrates, most of which are reviewed here. I would draw your attention to two in particular. Firstly, in a high profile paper in the journal Nature, Hallmann et al. (2014) demonstrate that geographic patterns of declines of insect-eating birds in the Netherlands are explained by local levels of environmental pollution with the neonicotinoid imidacloprid.  They conclude that this is unlikely to be due to direct toxicity, but that it is probably simply that in areas where insecticide use is high there are few insects for the birds to eat. Seems pretty plausible to me.

Secondly, a recent study of bird poisoning incidents in France shows that, between 1995 and 2014, there were 101 incidents involving 734 dead animals in which toxicological reports detected imidacloprid, and where the death is likely to be associated with agricultural use of seed treatments. 70% of these incidents happened during autumn cereal sowing. Grey partridges and various pigeon species were the ones most frequently killed.

According to Dr Hartfield, there have been no such incidents in the UK. This strikes me as very odd. Farming in France is very similar to the UK; most of the same crops are grown, and pesticide uses are very similar. I rang the UK Wildlife Incident Investigation Scheme to ask them whether they actually test dead birds for neonicotinoids. The person I spoke to did not know, and asked me to send in an email request, which I have done. When I get a reply, I’ll update the blog.

So, Dr Hartfield, my theories are more than “just theories”. “Real life” evidence from the Netherlands and France strongly supports them. The absence of direct evidence from the UK should not be interpreted as evidence of absence – if you don’t look for something, you won’t find it.   

The ongoing complacency of NFU, and their willingness to aggressively attack scientists who speak out against overuse of pesticides does them no credit. Farming bird populations are collapsing, along with most other farmland wildlife. Isn’t it time the NFU faced up to this, and took some responsibility?


Farmland bird population change (from JNCC)



Goulson, D. 2013. An overview of the environmental risks posed by neonicotinoid insecticides. JOURNAL OF APPLIED ECOLOGY 50: 977-987.


Are robotic bees the future?



Robot bee


There have been a number of scientific papers published in recent years discussing the possibility of building miniature flying robots to replace bees and pollinate crops. Clumsy prototypes have been tested, and seem to crudely work. If crops could be pollinated this way, farmers wouldn’t have to worry about harming bees with their insecticides. With wild bee populations in decline, perhaps these tiny robots are the answer?

While I can see the intellectual interest in trying to create robotic bees, I would argue that it is exceedingly unlikely that we could ever produce something as cheap or as effective as bees themselves. Bees have been around and pollinating flowers for more than 120 million years; they have evolved to become very good at it. It is remarkable hubris to think that we can improve on that. Consider just the numbers; there are roughly 80 million honeybee hives in the world, each containing perhaps 40,000 bees through the spring and summer. That adds up to 3.2 trillion bees. They feed themselves for free, breed for free, and even give us honey as a bonus. What would the cost be of replacing them with robots? Even if the robots could be built, complete with power pack and control devices, for one penny each (which seems absurdly optimistic) it would cost £32 billion to build them. And how long would they last? Some would malfunction, some would get caught out in the rain, some would be damaged by wind or spiders’ webs. If we very optimistically calculate the lifespan at one year, that means spending £32 billion every year (and continually littering the environment with trillions of tiny robots, unless they could be made biodegradable). What about the environmental costs of manufacture? What resources would they require, what carbon footprint would they have? Real bees avoid all of these issues; they are self-replicating, self-powering, and essentially carbon neutral.    

Thus far I have glossed over a vital further point. Pollination is not all done by honeybees. Numerous other insects pollinate crops and wildflowers, including butterflies, beetles, moths, flies, wasps and many more. These come in all sorts of different shapes and sizes suited to different flowers. Honeybees contribute at best 1/3 of crop pollination, averaged across crops. So we wouldn’t just need to replace the 3.2 trillion honeybees. We’d also need to replace countless trillions of other insects. All to replace creatures that currently deliver pollination for free.

Declines of bees are symptomatic of larger issues. It is not just bees that are declining; almost all wildlife is declining in the face of massive habitat loss and pollution across the globe. Even supposing we could create robot bees cheaply enough for it to be viable, should we? If farmers no longer need to worry about harming bees they could perhaps spray more pesticides, but there are many other beneficial creatures that live in farmland that would be harmed; ladybirds, hoverflies and wasps that attack crop pests, worms, dung beetles and millipedes that help recycle nutrients and keep the soil healthy, and many more. Are we going to make robotic worms and ladybirds too? What kind of world would we end up with?

Do we have to always look for a technical solution to the problems that we create, when a simple, natural solution is staring us in the face? We have wonderfully efficient pollinators already, let’s look after them, not plan for their demise. 

On neonicotinoids and impartiality in scientific research.



Last week my group posted online a lengthy review of new evidence (post 2013) relating to the possible risks posed by neonicotinoid insecticides to the environment: 

“The Environmental Risks of Neonicotinoid Pesticides: a Review of the Evidence Post-2013”

By Thomas James Wood, Dave Goulson

doi: https://doi.org/10.1101/098897

Free to view here: http://biorxiv.org/content/early/2017/01/06/098897

Funded by Greenpeace.

The rationale for compiling this review was that, in 2013, the European Food Standards Agency (EFSA) published a series of reports on three neonicotinoids (imidacloprid, thiamethoxam and clothianidin) which concluded that these chemicals pose an “unacceptable risk to bees”.  This conclusion led to the current EU moratorium which prevents the use of these three chemicals on flowering crops that bees might visit. EFSA are currently re-evaluating the position in the light of new scientific studies, and are expected to report back soon to the EU with a recommendation as to whether the moratorium should be rescinded, extended indefinitely in its current form or broadened to cover other uses (e.g. on non-flowering crops). It has been suggested that EFSA are compromised by close links with the agrochemical industry (though I have seen no evidence of this). Certainly industry are lobbying strongly to overturn the moratorium.  

Greenpeace asked us to write our own, independent review of the new evidence. Our review will make an interesting comparison with that produced by EFSA (now expected September 2017). Greenpeace paid the salary of my just-finishing PhD student, Tom Wood, for 4 months to complete this work with my help. For the record, other than Tom’s salary, they have not given any money to me or my lab, either recently or in the past. However, this does raise interesting issues about impartiality. I have previously pointed out that studies funded by, or carried out by, the agrochemical industry almost never find that neonicotinoids harm bees, while independent studies very often do find strong evidence of harm. This worries me because science depends completely on the honesty and impartiality of the experimenter. It is all too easy to bias results, consciously or subconsciously, in either direction. When patterns like this emerge, what is the lay-person to make of it? Sadly, it undermines public trust in science, at a time when good scientific evidence, and trust in this evidence, is urgently needed if we are to find a way through the myriad of terrifying environmental problems we are creating (climate change being an obvious example).     

So when we were offered money by a campaign group to write this review, we had to think very carefully. Greenpeace gave us written assurances that they would not attempt to influence what we wrote in any way, and they did not. Nonetheless, I am sure that some will say that this is just as flawed as industry evaluating the safety of their own chemicals. Of course there is one fundamental difference. When an agrochemical company conducts research on the safety of its own chemicals, it has a huge incentive to find a negative result. Pesticide sales are worth (literally) many billions of dollars per year. On the other hand, Tom and I get no direct benefit from publishing articles about pesticides, whatever we say. We might be more likely to get funding from Greenpeace in the future if they like what we write, so perhaps we are compromised to some extent too. On the other hand, if we published research that concluded that neonicotinoids were harmless we might be more likely to get funding from industry (though I think I may have burned that bridge long ago!), and industry has an awful lot more money than Greenpeace. So it is not clear that we have an obvious incentive to be biased, consciously or subconsciously, in one way or the other.

Nevertheless, I worry about impartiality and objectivity every day. I ask myself whether I am analysing patterns in the data correctly, or am I looking to confirm my own subconscious biases? When we conduct experiments in my lab, we try to report both negative and positive results as best we can (though journal editors don’t much like negative results). We recently published a paper which showed that low doses of the neonicotinoid clothianidin did not seem to have a measureable effect on learning or egg-laying in worker bumblebees (https://peerj.com/articles/1808.pdf ). If I had some sort of strange personal agenda to get neonicotinoids banned why would I publish this? With over 260 scientific publications to my name, I am not in desperate need of one more, and it might have been convenient to let this one slide, but we did not. We even paid the journal fees to get it published.   

Having spent six years studying neonicotinoids, I am pretty convinced that they are contributing to wildlife declines, not just those of bees. Hundreds of scientific studies from a huge diversity of different scientists on different continents conclude that they are harming bees, butterflies, aquatic insects, even birds and perhaps now bats (while a few studies found no effects). The balance of evidence seems to me to be overwhelmingly in one direction, despite industry efforts to obfuscate. That is my opinion, and it is not for sale. Unavoidably, all scientists in any field have opinions. Does this make us bad scientists? I hope not.

Of course, in an ideal world all research of this type would be funded entirely independently of stakeholders on either side. We don’t live in that world. So, if you have time, read the review. Make up your own mind whether it is fair or not.

The review contains no new data; it is simply a summary of what has been published in the scientific literature in the last four years. We have taken the unusual step of posting it on Biorxiv so that others can comment. This review has not yet been formally published; it is a work in progress. We encourage any interested party to give their opinion: have we missed critical information, have we wrongly interpreted studies? Once we have received comments, we hope to submit a final version for formal publication. This is not a procedure I have ever followed before, but seems to be a growing trend in this age of rapid, digital, open-access publishing.

So, comments please….