Field Tests For Pesticides:
How Robust Are They And Do They Protect Bees?

In the UK, we are constantly told that field studies for pesticides offer the favoured and realistic method of assessing risk of pesticides to bees. In contrast, doubt is often cast over independent laboratory studies.

On 13 September, Defra published an analysis of the results of its review of research done earlier in the year on the effects of neonicotinoid pesticides on bees. DEFRA is the UK Department For Environment, Food And Rural Affairs. It relies on various departments to conduct different aspects of services that inform and support DEFRA, and two of these departments include FERA (Food And Environment Research Agency) and CRD (Chemical Regulations Directorate).

I read the report, and it was clear there was a regular theme running through the comments that CRD made with regard to the independent scientific studies, some of which were laboratory studies. I have included them on the right.

Can you spot the recurring theme?

Well, it seems at least Bayer Cropscience would agree with CRD. They also seem very keen to suggest that what is important is ‘What’s happening in the field” as demonstrated in their response to the Pettis study (below).

“I am sure there are some very interesting effects Dr Pettis has seen in the laboratory, but in reality, when you get to what’s important to everybody, which is what happens in the field, you don’t see these things happening.” – Dr Julian Little, Bayer CropScience; source: The Independent: 20/01/2011, Michael McCarthy).

So one wonders:

"just how good are regulatory field tests for pesticides, and do they protect our bees?"

Aside from the fact that some of the studies featured right, were at least partial field studies anyway, nowhere that I am aware of, do FERA/CRD/DEFRA justify:

  • Why field tests should be better than laboratory studies, especially given all the potentially uncontrollable variables in field tests.
  • How field tests for pesticides could, in a scientifically ROBUST manner, measure many of the potential effects being investigated by the laboratory studies, such as behavioural effects on bees.
  • What it is that makes the regulatory field tests for pesticides so reliable and robust, such that we can trust them above the independent data implicating that pesticides are harmful to bees.

…Afterall, their report also states:

“The regulatory field studies fully comply with current guidance and also cover some additional aspects, such as over-wintering. The power of the studies to detect statistically significant changes is not established and they would not show all of the specific sub-lethal effects suggested by academic studies. However, hives exposed to treated crops did not show any gross effects on a wide range of important endpoints when compared to control hives exposed to untreated crops”.

So How Robust Are Regulatory Field Tests For Pesticides?

Well, EFSA, the European Food Standards Authority, despite various controverial influences from industry, conducted a review: Scientific Opinion on the science behind the development of a risk assessment of Plant Protection Products on bees (Apis mellifera, Bombus spp. and solitary bees), and they have found many MAJOR weaknesses with the regulatory standards for the field tests for pesticides.

(find here - opens new window).

Weaknesses include, funnily enough, unrealistic field conditions, short duration of experiments, tiny fields (taking into account realistic foraging areas for bees), wishy washy flexibility regarding important elements of the tests, inability for the field tests to measure impacts on colony, behavioural effects, chronic effects and so on. I could go on and on. Check out the report for yourself.

But this did not surprise me, having read the guidelines myself. The field test guidelines EPPO170 are devised, it appears, by a group of pesticide manufacturers and those related to the industry (e.g. providing consultancy services), as well as a sprinkling of government civil servants, such as Helen Thompson and Selwyn Wilkins of FERA.

(see work of the EPPO - opens new window).

Note the quote:

"At the ICPBR- Bee Protection Group 10th Symposium (Bucharest, 2009-10-08/10.....the WGs (working groups) presented proposals for the revision of EPPO Standards, which were discussed in order to hear the expert comments and recommendations of all 79 participants and to reach a consensus."

So who were the 79 participants? They were mostly industry related people.....

(see pages 158 -159 - opens new window).

“Field testing currently follows only one test guideline, the EPPO 170 (4). A precise analysis of this guideline, based on scientific knowledge, showed that it has several major weaknesses (e.g. the small size of the colonies, the very small distance between the hives and the treated field, the very low surface of the test field), leading to uncertainties concerning the real exposure of the honey bees. Even when they are carried out in natural conditions, they may not represent the normal exposure conditions of the bee colonies. Furthermore this guideline is better suited to the assessment of spray products than to that of SSST. For the moment, it does not allow us to assess all the effects of pesticides on the scale of the colony, including all bee categories and relevant long-term effects. It is therefore important and urgent to adapt guidelines to a specific mode of pesticide application and to the relevant exposure routes” we have an example of a field test conforming to EPPO 170 standards?

Unfortunately, I don’t have sight of a manufacturer’s regulatory field test for pesticides, but there have been scandals in the USA over Clothianidin, relating to poor standard field study data, and we know that our regulatory field study guidelines are lax.

But we do have a report written by FERA on behalf of DEFRA. It can be found on

(this link - opens new window).

It is a report that was produced in reaction to the concerns following a study by Girolami et al. (This study suggested that if bees consumed guttation drops (excretions of xylem fluid) from corn treated with neonicotinoids), they would die in minutes. Indeed, bees have been found to consume guttation drops. German beekeepers even investigated this themselves). Please note that neonicotinoids are 'systemic', meaning they permeate the whole plant and leaves - they are designed this way in order to kill insects that may chew on the leaves and suck on the sap.

Girolami et al: Translocation of Neonicotinoid Insecticides From Coated Seeds to Seedling Guttation Drops: A Novel Way of Intoxication for Bees. Journal of Economic Entomology 2009, 102 (5), 1808-1815).

In the report, FERA outline 3 different studies:

- Girolami’s.

- A Swiss government field study - one that conforms to EPPO 170 standards, it appears,

- A study by Shawkti looking at guttation from plants treated with a pesticide that was NOT systemic!

FERA’s treatment of the Girolami study is in contrast to their treatment of the Swiss study which is a deeply flawed as you'll see...... but first...

FERA MISLEAD the reader about independent data - this is AGAINST the Civil Servant's Code Of Practice!

FERA state, with regard to guttation:

“Only one study (Girolami et al 2009) has shown a significant effect in honeybees but this should be treated with caution as the data were generated by feeding collected droplets directly to bees and in many cases sucrose was added to ensure the honeybees consumed the dose.”


  • In FERA’s sentence above, it is made to sound that the addition of sucrose to all of the guttation fluid combined with the direct feeding to bees, is reason to be suspicious of Girolami’s study – almost giving the impression the bees are deliberately tricked into consuming pesticide. Firstly, this is untrue. In actual fact, Girolami added 15% honey only to some of the samples, and other bees were fed plain guttation drops. The addition of honey or not made no difference to the toxic effects - this significant point is not mentioned above.
  • In addition to which, regulatory tests for oral toxicity use a VERY similar method to that which they describe, for measuring acute oral toxicity in the laboratory! This is confirmed in EPPO 170 for Oral Toxicity tests. The test pesticides are mixed with a sucrose solution, and it is fed directly to bees via single use feeders. FERA know all about regulatory test methodologies and indeed, Helen Thompson of FERA certainly knows this – she conducted the oral toxicity tests on behalf of Bayer CropScience to support the DAR for imidacloprid. In these test, she fed bees test solution mixed with sucrose, as described.

It's unfortunate that many will be misled by FERA's summary of Girolami's study - which in itself has to be read in its ENTIRETY (and probably a couple of times) to be properly grasped.

The Swiss study was called: “Presence of Clothianidin in Hives (Monitoring)”, and the objective of this study was “Quantitatively and qualitatively establish the presence of neonicotinoids (clothianidin)”

The field study investigated clothianidin residues in guttation, pollen, honey and dead bees. This is FERA's full summary from document:

“The Swiss Federal Government for Agriculture commissioned a study in 2009 to assess the risks to honeybee colonies during sowing of maize seed treated with Poncho (25g ai/ 50,000 seeds, i.e. 0.5 mg ai/seed) through drift of dust and guttation. No effects were observed due to dust drift. Guttation fluid collected from maize after emergence (7-10 days after sowing) was reported to contain 25-37 mg clothianidin/L reducing to around 0.1 mg/L by 40 days after sowing (as above the LD50 for clothianidin is around 0.084 mg/L) (Figure 3). No clothianidin residues were detected in the honeybees or in honey sampled from the colonies and no increased mortality was identified at honeybee colonies placed at the edges of the treated fields and the colonies developed normally”.

Well it's interesting. In this deeply flawed experiment, the only part that could be considered a little robust was the measuring of clothianidin in the guttation samples - of which the authors found high concentrations. FERA do not draw attention to the fact that the guttation levels are considered high - and the figures will mean little to those 'not in the know'.

And was the experiment appropriately designed to "assess risks to honeybee colonies" as FERA state above? No, but FERA omit to mention this.

When you follow the link to the Swiss study you'll note it is written in French, (
hence I have summarised the study for you.

There were 2 elements to the test.

Test 1

- A 2 hectare field was used – this is TINY - foraging bees will fly several kilometres – EFSA state that even at only 2.5 km - a realtively short flying distance, this corresponds to a theoretical foraging area of 19.6 km2, i.e., about 2000 hectares, in contrast to the 2ha used in the study.

- Only 36% of this tiny field was sown with clothianidin, the pesticide that was the subject of the study!

Given that FERA keep wittering about ‘field realistic conditions’, can anyone let me know of any modern agricultural crop field, where not only is the field itself tiny, but only about a third of it is actually sown with the crop? Because I don’t know of such a field! There is no criticism by FERA of the study with this regard.  And please remember that field studies fail to take account of bees not returning to the hive (perhaps due to death or disorientation, for example).

So what they are saying is, 74% of this tiny field had no clothianidin at all - the very substance they were testing for!

Well, to be fair, the EPPO 170 guidelines state the field should be at least 1 ha in size, but they don’t stipulate it has to be sown completely with the pesticide being tested. It turns out that what was treated with clothianidin in reality, were 8 rows of corn, that were by the side of wildflower strips, close to apple trees and dandelions.

The rest of the field was sown with:

  • Gaucho - imidacloprid
  • Cruiser - thiamethoxam
  • Mesorol - which is not a systemic neoniocotinoid. It is a bird repellent and molluscicide and a non-systemic carbamate pesticide. This is relevant, because the point about systemic pesticides like neonicotinoids, is that they permeate the whole plant, hence the risk through the guttation drops. It is possible this was therefore the control substance, but it is not clear.

The author does not state what proportion of the field was treated each with Mesorol, Gaucho and Cruiser. The study does not even refer to these other pesticides later – so we don’t know anything further about residues of other neonicotinoids. FERA do not draw attention to this point either.

Anyway, colonies with 20,000 bees each were installed by the side of the clothianidin treated field (with just 8 treated corn rows by the side of wildflower strips). 6 colonies of bees were used on this tiny field – 3 on each side – in reality, these colonies must have been close together – a regulatory field trial fault also identified by EFSA.

- 4 colonies from the sample of 6 (2 from each side) were used for monitoring of dead bees collected in traps at the hive. Pollen and honey samples were collected from the remaining 2 hives and were analysed for residues of clothianidin. In the case of the honey, the samples were taken 3-5 weeks before and after planting the corn.


This trial would not be able to take account of bees that did not return to the hive, due to disorientation or death from pesticide. 

- The trial began 17th April – the seeds were sown April 23rd. The trial lasted ONLY 50 days from April 23rd (where April 23rd = 0 in the summary table). 50 days (also includes the time taken for the corn seedlings to appear) is certainly insufficient time to see:

  • colony effects,
  • chronic effects
  • behavioural effects on the bees

- even if the field conditions had been sufficiently realistic to gather robust data (taking into account that Bayer CropScience in their leaflet “The Secret Life Of Termites” (which are colony insects like honey bees) refer to a time period of 3 months to kill a colony).

- Guttation was collected 14 times and analysed in the lab for clothianidin.

Test 2

The second field test was only 200m to 300m from the first field test. The field was surrounded by wildflowers, dandelions, clover. The bee colonies were placed on either side of the 1 hectare field in a meadow, onto which a strip was built (what size?) and upon which, seeds treated with Poncho (clothianidin) were sown using a mechanical seeder. 3 honey bee colonies were placed on either side of the meadow 16 days before planting time. During this time, the bees would have been busy foraging elsewhere, building stocks in the hive. The rest of the field study is of the same standard of ‘science’ (if you can call it that) as above. Clothianidin residues were measured, with predicable results.


The authors find "no significant mortality" of bees (although disorientation effects and deaths of bees outside of the hive due to poisoning would not of course, be recorded with this flawed field study in any case). They find, however, 10 dead bees per day (in a bumble colony, losses of 10 bees per day would be devastating, since a colony may typically only consist of about 160 bumblebees).

They do not detect residues of clothianidin in pollen and honey, with the exception of one sample of pollen (probably because the bees were foraging on flowers flowers they prefer). But they do detect contamination in ONE sample. The authors state this is likely to be caused by increased exposure to clothianidin of the dandelions close by, due to frequent use of the field for trials. That's an interesting admission, that could have several implications:

  • the authors of the study know full well that the bees were unlikely to forage on the corn, given the tiny size of the field with more attractive foraging opportunities close by, however in monoculture situations bees may have little choice but to forage on the treated crop, whatever that is;
  • field margins are not safe from contamination (confirmed in other studies already - e.g. Tapparo et al);
  • if those trials had been conducted in the past, it suggests the dangers of contamination from residues that are retaken from soil contaminated from previous sewings, taking into account that neonicotinoids are persistent in soil.

The author find levels of clothianidin in guttation samples from 27 to 37,000 (thirty seven thousand) microlitres per litre. Even the authors (using the debatable LD50 measure) have to concede that this is high....

........but since during their unrealistic experiment they haven’t found any clothianidin residues in dead bees, honey or pollen during the time period the corn was guttating, the authors conclude this must mean the guttation doesn’t pose a threat. But they advise that if water is of limited availability to the bees, then beekeepers should engage in good beekeeping practice, by supplying water in the beehive.

Could some-one also ask them to do the same for the wild bees and invertebrates at risk as well, please?

Within their discussion they refer to bee poisoning cases in Germany, and state that no such observations have been recorded in this experiment. Could that be because the German situation reflected reality, where as this unrealistic Swiss Government study does not? I suspect this is the case.

More interesting interpretation from FERA.....

FERA go on to summarise a further study by Shawki et al (2006). It should be noted that this study does not add any weight at all to the notion that guttation from crops treated with systemic insecticides, is not harmful to bees. Nor does it support in any way, the findings of the Swiss study above. The fact that the crops were treated with a non-systemic insecticide seems highly relevant. This is not explained by FERA, nor is it properly explained why guttation results would be so different from those in the other studies. As such, FERA are omitting important information that will potentially mislead the reader of this report.

FERA use a poor quality field study to cast doubt on other data...

Fera: "However, as studies in Switzerland showed no significant mortality in bee colonies located at the edge of treated maize fields the significance of guttation fluid as a source of water for bees is unclear”.

FERA refer to the unrealistic Swiss study again (right), to cast doubt on other evidence suggesting that guttation could be highly toxic for bees.

So they used an unrealistic and dodgy study in Switzerland to cast doubt on the credible, independent data on guttation..........)

And so what can we conclude from all this? How about....

  • That studies conforming to regulatory field tests for pesticides won't produce robust data - so how could such studies support the claim "there is no unacceptable risk to bees".
  • Our bees and other invertebrates are not protected by our regulatory system - which means the environment as a whole is not protected, so neither are food supplies nor ultimately, the well-being of humanity.
  • The behaviour of our civil servants is disgusting. They require 'trash science' to support registrations of poisons, and 'unequivocal evidence' to dislodge them.
  • That our Civil Servants mislead us - this is against the Civil Servant's Code of Practice in the UK.

    .............I can think of a few more....but I'll let you think of a few yourself!

Go to

FERA betray bees again

National Bee Unit And Nosema

Bees And Neonicotinoids

Go from Field Tests For Pesticides back to Home page

Protected by Copyscape DMCA Takedown Notice Checker



Book Review

Click Here

CRD Response to Independent Tests:

Henry et al: A common pesticide decreases foraging success and survival in honey bees

“Due to the artificiality of the test design and dosing regime, there are uncertainties regarding the risk in a more realistic field exposure situation”.

Whitehorn et al: Neonicotinoid pesticide reduces bumble bee colony growth and queen production

“The key question for this study is how far it illuminates the likely real situation at field level - are the exposure and the resulting effects seen under normal conditions?”

Pettis et al: Pesticide exposure in honey bees results in increased levels of the gut pathogen Nosema

“The following points require consideration in trying to interpret this study: whether factors such as exposure are in line with field situations”.

Vidau et al: Exposure to sub-lethal doses of fipronil and thiacloprid highly increases mortality of honey bees previously infected by Nosema ceranae

“The key issue is how this relates to potential exposure under field conditions – i.e. will bees be exposed to this level of pesticide under field conditions?”

Wu et al: Honey bees (Apis mellifera) reared in brood combs containing high levels of pesticide residues exhibit increased susceptibility to Nosema (Microsporidia) infection

“There is no information regarding the relevance of the concentrations given to the bees – i.e. were the levels appropriate and/or realistic?”

Mommaerts et al: Risk assessment for side-effects of neonicotinoids against bumblebees with and without impairing foraging behaviour

"It is unclear as to the relevance of the concentrations tested and hence whether the results can be extrapolated to the field situation”.

Johnson et al: Ecological appropriate xenobiotics induce cytochrome P450s in Apis mellifera

“the issue of detoxification and the relevance of diet should be considered under realistic field conditions.”

Schneider et al: RFID Tracking of Sub-lethal Effects of Two Neonicotinoid Insecticides on the Foraging Behavior of Apis mellifera

“the next logical step would be to employ this methodology at the field scale to see if the effects observed in this study were replicated under field conditions”.

Teeters et al: Using video-tracking to assess sub-lethal effects of pesticides on honey bees (Apis mellifera L.)

“As regards the effects seen, it is important to note that they need to be related to what is likely to be encountered either in the hive (i.e. were the rates used realistic in terms of exposure to tau-fluvalinate) or in the field (i.e. were the rates used realistic in terms of what a worker honey bee is likely to encounter in the field.)”

Dr Chris Connolly,
University of Dundee (commenting on the
Sept 13th report):

“With respect to the criticism of the validity of all lab studies, past and future, in preference for the more relevant field studies, I consider this claim totally unprofessional and lacking all scientific credibility. Laboratory studies are the cornerstone of all therapeutic drug discovery as they provide a mechanistic description of events that can be controlled and tested experimentally. These studies identify real and quantified threats. In contrast, field studies are performed in a particular context with an uncontrolled surrounding area. What may be found at one site could be irrelevant to that found at another site. This is especially important given the multiple stresses to which our pollinators are exposed and the likelihood that multiple threats contribute to the pollinator decline. It is true that a laboratory based mechanistic explanation does not confirm that these effects are largely responsible for the pollinator decline. This will require countrywide bioinformatics once we know what pesticides have been used. An isolated field study has limited value”.