Late Lessons, Jensen Huang and AI
Markdown version

Hindsight check: LL2-16 (Ch 16 Seed-dressing systemic insecticides and honeybees)#

Source section: EEA, Late lessons from early warnings: science, precaution, innovation (EEA Report No 1/2013), Ch 16, by Laura Maxim and Jeroen van der Sluijs (report pp. 369–406; PDF pp. 371–408). The chapter includes Panel 16.1, the Bayer CropScience view by Richard Schmuck (pp. 401–402), and Panel 16.2, the authors’ reply (pp. 403–406). Check window: publication (2013) to late September 2026. The chapter dates itself “at present, February 2012” (p. 374). Checked: 25 September 2026.

Method note. - Search. General web search was unavailable for this pass because the session’s search budget was exhausted. I retrieved sources directly from primary repositories: - EUR-Lex and the EU Publications Office (the regulations, directives and CJEU judgments, all in full text); - EFSA’s website and EFSA Journal full texts in PubMed Central; - DG SANTE web pages and the summary reports and agendas of the Standing Committee on Plants, Animals, Food and Feed (SCoPAFF), Phytopharmaceuticals–Legislation section, 2023–2026; - the European Court of Auditors; - the French Conseil constitutionnel (decisions and government observations); - Légifrance; - Europe PMC, for peer-reviewed abstracts and, where open access, full text. - Access gaps. - The EFSA Journal pages on Wiley sit behind a bot check, which I did not try to get around. For EFSA’s January 2013 conclusions I therefore relied on EFSA’s own press release and the recitals of Regulation 485/2013. - The French case-law databases (Judilibre, Légifrance) need JavaScript. For the outcome of the criminal case (Claim 10) and the June 2025 Conseil d’État ruling, my only source is the website of UNAF, the beekeepers’ union and a party to the case. I say so where it matters. - I did not retrieve French national series on colony mortality or honey production after 2004. - Abstracts. Where a point rests on an abstract only, I say so. - Annex 3 does not apply. This is a new 2013 case with no 2001 counterpart. - Scope. I read only the section digest, the section’s source extract and web sources. Page numbers are report pages of the 2013 volume. - Conflicts of role. - The chapter’s authors and their network. Van der Sluijs is lead author of the 2015 “Worldwide Integrated Assessment” (WIA) conclusions of the Task Force on Systemic Pesticides [S72]. Bonmatin, whose studies feature heavily in the chapter, co-authored the WIA and a 2024 French hazard-trend study [S57]. Other WIA authors (Goulson, Sánchez-Bayo) co-wrote the 2016 pesticide–disease review [S46]. The EU’s 2023 residue regulation cites the WIA in its recitals [S19]. So some of the later evidence and its uptake in policy comes from the chapter authors’ own network. - Industry funding. Rolke et al. 2016 was funded by Bayer CropScience [S42]. Woodcock et al. 2017 was co-funded by Bayer CropScience and Syngenta, with NERC funding (Europe PMC grant metadata) [S37]. - Parties to disputes. UNAF was a party to the criminal case and to the EU litigation.

Overview#

1. The chapter’s central diagnosis about method was confirmed almost at once. It has since been written into guidance, but still not put into routine use. - The chapter said the acute, spray-based hazard-quotient (HQ) scheme was unfit for systemic seed-dressings (pp. 377, 389–390). - Who agreed: - EFSA’s 2012 opinion [S3]; - EFSA’s 2013 Bee Guidance Document [S4]; - the EU courts [S17]; - the European Court of Auditors, which in 2020 called the EU scheme “unsuitable” for products applied to seeds or soil, “such as neonicotinoids” [S1]. - What held up implementation: - From 2013 a majority of Member States refused to endorse the 2013 guidance, especially its chronic-toxicity parts [S1, S2]. - The Commission’s 2002 guidance remained the basis for approvals [S2]. - EFSA published a revised guidance in May 2023. It sets a looser protection goal for honeybees than 2013: at most a 10% reduction in colony size, against 7% [S5, S1]. - The regulations needed to apply it were still awaiting a committee vote scheduled for 24–25 September 2026 [S6]. - An EFSA conclusion published in May 2026 still ran the bee assessment under the 2002 guidance and recorded a “data gap for sublethal effects” [S7].

2. The case against imidacloprid seed-dressing grew stronger and became EU law. - The steps: - EFSA’s January 2013 conclusions [S8]; - restrictions on bee-attractive crops in 2013 [S10]; - an EFSA update in February 2018 finding risk to all three bee groups for most uses [S12, S13]; - a ban on all outdoor uses in 2018 [S14]; - lapse of the approval on 1 December 2020 [S15]. - The courts. The EU courts upheld the 2013 restrictions in 2018 and 2021, explicitly accepting precautionary action under scientific uncertainty [S16, S17].

3. The claims on persistence and unexpected exposure routes are among the best confirmed in the chapter. - Residues outside treated crops. Neonicotinoids were found in: - wildflowers in field margins, sometimes at higher levels than in the crop [S51]; - pollen from non-crop plants throughout the season [S52]; - soils and plants on organic farms [S54]; - 75% of honey samples worldwide [S55]. - Succeeding crops. EFSA itself assessed exposure in succeeding crops [S13]. - Persistence after the moratorium. In western France, imidacloprid residues in oilseed-rape nectar showed no clear decline across the five years of the 2013 partial moratorium [S53]. - Surface waters. In France the risk from imidacloprid in surface waters fell sharply only after the 2018 total ban [S57]. The EU made imidacloprid and four other neonicotinoids priority water pollutants in 2026 [S24].

4. Two of the chapter’s scientific claims are weaker than presented. - Synergy with the Nosema pathogen. Table 16.1 (p. 394) says this was “proving[d]”. Later work contests it: - a critical review found no pesticide–pathogen interaction on survival under field-realistic conditions [S45]; - a 2021 meta-analysis found pesticide–parasite interactions on mortality to be additive, not synergistic [S43]; - free-flying and field studies found no synergy [S49, S41]. - The honeybee as a sensitive sentinel (p. 393). Field studies found honeybee colonies relatively robust, while wild bees were harmed [S36, S37, S41, S44]. - Upshot. The harms that did hold up best lie beyond the honeybee: - wild bees [S36, S39, S44]; - aquatic invertebrates [S61, S24]; - insectivorous birds [S62, S58]. The chapter was right that the problem reached “other entomofauna”. But the visible honeybee losses that set off the controversy were the hardest harm to prove at colony level.

5. “The bans worked” rests on reduced exposure, not on shown recovery. - Exposure did fall: - in rural areas of the UK [S56]; - in French surface waters [S57]. - No study measured the effect on colonies. I found no independent evaluation of how the French bans affected colony mortality. - Recovery in birds was weak. In France, insectivorous birds showed only weak recovery after the 2018 ban [S58]. - Colony losses remain multi-causal, in the pan-European EPILOBEE surveillance study and in Italy [S65, S66]. - Bayer’s “no correlation” claim was contradicted at national scale for England and Wales [S40]. It was supported for honeybee colonies in some field studies [S41, S42].

6. Costs and lock-in: the chapter was right about preventive use, but its national-yield argument does not generalise. - Preventive overuse is confirmed. In several crops, seed treatment brought little yield benefit [S67, S68, S69]. - Sugar beet is the counter-case. After France’s 2018 ban, French beet yields fell from 851 to 649 quintals per hectare (q/ha) between 2019 and 2020, during an outbreak of virus yellows (an aphid-spread disease) [S28]. - A cycle of derogations and substitutes followed: - 206 EU emergency authorisations for the three restricted neonicotinoids in 2013–2019 [S1]; - French beet derogations from 2021, under a law upheld in December 2020 [S28, S57]; - a CJEU ruling in January 2023 that emergency authorisations cannot override the treated-seed ban [S18]; - a 2025 reauthorisation attempt struck down by the Conseil constitutionnel [S29]; - time-limited derogations in August 2026 for flupyradifurone (a substance with the same mode of action) and acetamiprid (another neonicotinoid), which the Conseil constitutionnel upheld with reservations [S30].

7. The governance recommendations were largely adopted, but mostly through other controversies. - What was adopted: - the EU Transparency Regulation (2019), with public studies, a register of commissioned studies and public declarations of interest [S21]; - the Whistleblower Directive (2019) [S22]; - the anti-SLAPP Directive (2024), which names researchers and academics [S23]; - a French law of April 2013 on independent expertise and whistleblowers [S31]. - Where the impetus came from. The Transparency Regulation cites the glyphosate citizens’ initiative, not this case [S21]. - A counter-current. A December 2025 Commission proposal would make most approvals unlimited in time [S26]. That would reduce the scheduled re-reviews through which new knowledge entered this case.

8. The criminal case ended without a trial. - The ruling. An examining judge dismissed the case (a non-lieu) on 1 April 2014, citing, as reported by UNAF, scientific uncertainty about whether bee losses could be attributed to Gaucho. - The appeals. The Paris Court of Appeal confirmed the dismissal on 15 April 2015, and the Cour de cassation (France’s highest court for criminal matters) rejected UNAF’s appeal on 4 January 2017 [S32, S33]. - What it shows. Administrative and constitutional judges accepted precautionary action, while criminal law demanded causal proof.

Implication for weight. - Give strong weight to the chapter’s lessons on mechanisms: - legacy assessment tools applied to a new mode of exposure create blind spots; - persistent, mobile agents reveal new exposure routes only after wide use; - demanding a single cause produces inaction; - restricting single products leads to substitution by near-equivalents and repeated derogations; - preventive, bundled use creates lock-in. Each is independently confirmed after 2013. - Add a lesson the chapter could not state in 2013: agreeing that the method was wrong did not mean the replacement method came into force. Thirteen years later it still awaits adoption. - Give moderate weight to the institutional narrative. Its later confirmations come partly from the authors’ own network. - Give low to moderate weight to specific ecotoxicological claims, namely: - Nosema synergy; - the honeybee colony as sentinel; - recovery of colonies after the bans; - the argument from national yields.

Claim-by-claim#

Claim 1: Standard sprayed-pesticide risk assessment (acute LD50, application-rate HQ) is inappropriate for systemic seed-dressings; new standardised tests for chronic, sublethal and colony-level effects are needed; the 2010 EPPO revision “failed to significantly change the risk assessment pattern”, so the Commission mandated EFSA to review it (pp. 377, 389–390)#

Original claim (pp. 377, 389–390). - The 2010 EPPO norm, the European and Mediterranean Plant Protection Organization (EPPO) standard, “still does not consider sublethal and chronic effects properly”. Standardised laboratory tests for chronic and sublethal effects “are still missing”, so the Commission mandated EFSA (p. 390). - Lessons 1 and 2 (pp. 389–390) call for checking whether methods fit a new technology and for developing tests for chronic, sublethal and colony-level effects. - The summary box (p. 369) already says these insights “were recently confirmed” by EFSA.

Subsequent developments - Diagnosis confirmed. - The Commission asked EFSA on 18 March 2011 to review the EPPO scheme for chronic risk, low-dose exposure, guttation and cumulative risk. EFSA’s opinion of 23 May 2012 identified “several weaknesses in the EPPO Guidance” (as recounted by the CJEU, paras 16 and 20) [S17, S3]. - EFSA’s Bee Guidance Document (4 July 2013, updated July 2014) did the following [S1, S4]: - added chronic tests for adults and larvae; - covered bumble bees and solitary bees; - added exposure through dust, guttation fluid and water. - The 2013 and 2018 neonicotinoid reviews used it [S2, S12]. - The European Court of Auditors (Special Report 15/2020, July 2020) described the scheme still in force as an HQ scheme with a trigger value of 50. Because it assumes spray exposure, it is “unsuitable for PPPs that are applied to soil or seeds (such as neonicotinoids)” (PPPs are plant protection products) [S1, Box 6]. - Implementation blocked. - Since 2013, 12 Member States supported the 2013 guidance and the rest “consistently objected” [S1, para 52]. - A SCoPAFF meeting in December 2013 gave three reasons it could not be fully applied [S1, paras 52–53]: - too few internationally agreed test methods; - protection goals seen as unrealistic; - impractical designs for field studies. - The Commission says the majority objected “in particular” to the chronic-toxicity parts [S2]. - In July 2019 Member States backed applying only the part on acute toxicity to honeybees. The European Parliament vetoed that in October 2019 because it wanted chronic toxicity included [S2, S1]. - The Commission states the 2002 guidance “remains the basis” for approvals and renewals [S2]. - Tests came slowly and were not required. - OECD adopted the following tests [S1, Annex I]: - larval tests (TG 237, 2013; Guidance Document 239, 2016); - a 10-day chronic adult test (TG 245, 2017); - acute tests for bumblebees (TG 246/247, 2017). - A homing-flight test for sublethal effects, led by France, was still in progress in 2020. - The Commission told the auditors it had “not yet required applicants to use test methods developed after 2013” because the guidance had not been endorsed [S1, Commission replies]. - The 2013 guidance itself said it was “not possible to consider sub-lethal effects” in its schemes [S1, Annex I]. - The 2023 revision. - EFSA published revised guidance on 11 May 2023 [S5]. It is tiered and covers acute and chronic effects, larvae, metabolites and mixtures. For honeybees it “considers possible long-term effects of low doses and potential concerns due to sublethal effects”. - Risk managers set the protection goal at no more than a 10% reduction in honeybee colony size. The 2013 figure was 7% [S1, S5]. At a February 2021 meeting of 23 Member States, the views were [S2]:

| Maximum acceptable reduction in colony size | Member States |
|---|---|
| Up to 23% | 4 |
| 10–12.8% | 11 |
| Keep 7% | 4 |
| No preference | 4 |

Verdict: strengthened. - The diagnosis was adopted by EFSA, the auditors and the courts. - The chapter’s own evidence of inertia (the EPPO revision “failed” to change the pattern) was also borne out. The replacement method has taken more than 13 years and is still not applied. - The protection goal was loosened along the way (from 7% to 10%).

Implications for weight. This is the chapter’s most robust lesson. It needs one addition. Even when experts agree a method is unfit, replacing it depends on: - agreement among risk managers on “acceptable” effect sizes; - test guidelines that exist and are required; - amending the legal data requirements.

Each of these is a separate veto point. Protection goals get renegotiated along the way.

Claim 2: The French expert committee (CST, 2003) found significant risks to honeybees from Gaucho in sunflower and, via pollen eaten by nurse bees, in maize; imidacloprid was nonetheless approved EU-wide in 2008 on a German draft assessment that omitted the French exposure studies (pp. 380–381, 388)#

Original claim (pp. 380–381, 388). - The CST (Comité Scientifique et Technique, the French expert committee) concluded in 2003 that seed-dressing posed significant risks, including to nurse bees in maize. - The German draft assessment report (DAR), submitted to EFSA in 2005, “accorded only limited importance to sublethal effects”. It omitted, for example, all of Bonmatin’s studies, and its method “was not adapted to seed-dressing formulations” (p. 388). - Imidacloprid was listed at EU level in 2008 (p. 388). EFSA gives the date of inclusion as 1 August 2009, by Directive 2008/116/EC [S13, Background].

Subsequent developments - EFSA, 16 January 2013 [S8, S9]: - For dust, a risk to honeybees “was indicated or could not be excluded, with some exceptions”. - For pollen and nectar, “only uses on crops not attractive to honey bees were considered acceptable”. - It flagged many data gaps and limited information on other pollinators. - I did not verify whether this EFSA review drew on the French studies the DAR omitted. The full text sits behind the bot check on Wiley. - Regulation (EU) No 485/2013 (24 May 2013) [S10]: - restricted clothianidin, thiamethoxam and imidacloprid on bee-attractive crops, which the Commission’s summary says include maize and oilseed rape [S15]; - banned the sale and use of treated seed from 1 December 2013; - recited “high acute risks” from dust, from pollen and nectar for some crops, and from guttation in maize. - Fipronil. Fipronil, co-used and co-banned in France in 2004, was restricted separately (Regulation 781/2013) because of dust risks in maize [S11]. This confounds any attempt to attribute outcomes to imidacloprid alone. - EFSA, 28 February 2018. The update, based on a systematic literature review and an open call for data, concluded that “most uses of neonicotinoid pesticides represent a risk to wild bees and honeybees” [S12]. - For imidacloprid, a “low risk was concluded for some crops for honeybees”. But when all three bee groups are considered, “a high risk was concluded or … a low risk was not demonstrated for all the uses assessed”. This applied both to pollen and nectar and to dust [S13, Summary]. - For guttation, the risk to honeybees was low for winter cereals, sugar beet and potatoes, and high for all other uses [S13]. - Regulation 2018/783 (29 May 2018) limited imidacloprid to permanent greenhouses [S14]. The approval then lapsed on 1 December 2020 when no renewal application was supported [S15]. - Courts. - The EU General Court dismissed Bayer’s and Syngenta’s actions against the 2013 restrictions on 17 May 2018 (T-429/13 and T-451/13) [S16]. - The CJEU dismissed Bayer’s appeal on 6 May 2021 (C-499/18 P). It held that “an exhaustive risk assessment cannot be required” where the precautionary principle applies, and that measures need not wait for ongoing studies (paras 79–82) [S17]. - The CJEU did find a legal error in how the General Court set the threshold for review under Article 21, but held that ground of appeal ineffective (paras 49, 56) [S17].

Verdict: strengthened. - The EU assessment converged on the French committee’s broad conclusion and went further: all outdoor uses, all bee groups. - The 2008 approval rested on a method later judged unfit [S1, S3]. - One caveat: the CST’s specific mechanism, chronic exposure of nurse bees through maize pollen, was not singled out. EFSA’s 2013 concerns centred on dust and data gaps. By 2018 some honeybee pollen risks were judged low at higher tiers, and the risks that were confirmed concerned bumble and solitary bees.

Implications for weight. Strong support for the lesson that selective use of evidence and an ill-fitting method at authorisation can be corrected, but late: 1994 first use, 2013 EU restriction, 2018 outdoor ban. The courts’ endorsement of acting under uncertainty is a significant later vindication of the chapter’s precautionary framing. A later review from science and technology studies treats the EU restrictions as an unusual case in which academic ecotoxicology, rather than standard regulatory science, drove regulation [S71] (abstract only). That fits the chapter’s account of public scientists contesting the dossier.

Claim 3: Imidacloprid persists and reaches bees by routes not anticipated at authorisation: soil half-lives of 188–249 days in the manufacturer’s dossier; 1–2 ppb in untreated succeeding crops; sowing dust, guttation and year-round pollen residues; a single flight through drill dust gave about 300 ng per bee (pp. 376, 387, 393, 404–405)#

Original claim. - Bayer’s dossier gave soil half-lives (DT50) of 188 ± 25 and 249 ± 40 days, above the three-month trigger in Directive 91/414 (p. 376). - Untreated sunflowers grown a year later carried 1–2 ppb (p. 387). - Other exposure routes: sowing dust, guttation and wildflowers coated with dust (p. 387). Residues were found in pollen all year and at all sites (p. 405). - A single flight through the dust cloud of a pneumatic drill delivered an average of 300 ng per bee, even with deflectors (p. 404).

Subsequent developments - Regulatory acceptance of these routes. - The 2013 guidance and the 2013 and 2018 reviews added dust, guttation, water and succeeding crops as exposure scenarios [S1, S4, S13]. - EFSA 2018 used the longest field DT50, 288 days, to model soil accumulation. It found “moderate evidence for larger than negligible effects” in the succeeding-crop scenario [S13]. - Wild plants and non-crop exposure. - UK field margins: neonicotinoids were in wildflower pollen and nectar, sometimes at higher levels than in the crop, and 97% of the neonicotinoid in hive pollen came from wildflowers [S51]. - US maize and soy landscapes: pollen from non-crop plants was contaminated all season. Pyrethroids, not neonicotinoids, were the highest residues [S52]. - Persistence despite restrictions. - Oilseed-rape nectar in western France, 2014–2018 (under the 2013 moratorium), contained imidacloprid every year with no clear decline. Levels depended on soil type and rainfall. The authors conclude that persistent soil residues “diffuse on a large scale” [S53] (abstract). - Swiss survey: neonicotinoids were found in 93% of organic soils and crops, and in over 80% of ecological focus areas [S54]. - Honey worldwide: 75% of 198 samples contained at least one neonicotinoid, at levels below human residue limits [S55]. - Rural UK bumblebee exposure fell after the moratorium; peri-urban exposure did not [S56]. - Water. - An EEA assessment of 2013–2019 found pesticide exceedances at 13–30% of surface-water sites each year, “mainly caused by” imidacloprid and malathion among insecticides (cited in the Commission’s 2022 impact assessment) [S25]. - In France, imidacloprid-related hazard in water rose until 2018 and “decreased sharply after 2018” [S57]. - Directive (EU) 2026/805 (30 March 2026) lists imidacloprid as a priority substance, with an annual-average environmental quality standard of 0.0068 µg/l for inland waters. It also lists acetamiprid, clothianidin, thiacloprid and thiamethoxam. Transposition is due by 21 December 2027 [S24]. - Dust. - Italian engineering trials reported that air deflectors achieved only an “unsatisfactory” 50% cut in active ingredient. In static tests, a filtering and recirculating prototype cut dust by 98% and active ingredient by 97% [S59] (abstract). - In US maize country, three years of monitoring found clear short-term harm to colonies at corn planting [S60]: - seed-treatment residues in pollen during planting; - elevated worker mortality. - EU regulators identified dust as a primary acute risk in 2013 [S10, S11].

Verdict: strengthened. Every exposure route the chapter listed was later confirmed and written into EU assessment practice. Persistence proved serious enough that residues lingered for years into the partial moratorium. They fell clearly only after the total ban.

Implications for weight. This is strong, independent support for the lesson that persistent, mobile agents reveal their exposure routes only after wide use. A later corollary: withdrawing such an agent reduces exposure slowly and unevenly, so a restriction’s benefits lag behind it.

Claim 4: Sublethal imidacloprid exposure acts synergistically with the pathogen Nosema, weakening colonies (p. 387; Table 16.1, p. 394, calls it “proving”)#

Original claim. - Alaux et al. (2009) found the highest mortality when bees were both infected with Nosema and exposed to imidacloprid, with long-term immunosuppression (p. 387). - Table 16.1 lists 2009 as the year of a “Scientific publication proving synergic effects between imidacloprid and Nosema” (p. 394). - The reply to Bayer cites four such studies (p. 404).

Subsequent developments - Supporting (mostly laboratory). - Clothianidin, and similarly imidacloprid, suppressed NF-κB immune signalling and promoted deformed-wing-virus replication at sublethal doses [S47]. - Thiamethoxam at low doses plus Nosema apis synergistically increased mortality in an Australian population [S48]. - A 2016 review, co-written by WIA authors, argued that immune suppression links pesticides to disease [S46]. - Varroa mites made thiamethoxam-induced homing failure worse [S75] (abstract). - Contrary or qualifying. - A critical review (2016) found pesticide exposure and pathogen infection “have not yet been found to interact to affect worker survival under field-realistic scenarios”. Colony-level effects remained “unclear” because the effects on Nosema had been shown only in the laboratory, generally at exposures above those likely in the field [S45]. - A meta-analysis (2021) of 356 interaction effect sizes found strong synergy among multiple agrochemicals. Interactions with parasites and nutritional stress were “not greater than additive” for mortality [S43]. - In free-flying mini-hives, clothianidin plus Nosema showed no synergy. The authors suggest “social buffering” within the colony [S49] (abstract). - Two field studies found no difference in pathogens: - Sweden, clothianidin, over two years: no detectable negative impact on colonies or pathogens [S41]; - Spain, thiamethoxam oilseed rape: pathogens did not differ significantly between exposed and control apiaries [S50].

Verdict: contested. - Laboratory interactions exist, and a plausible immune mechanism has been described. - The chapter’s “proving” overstated a single 2009 laboratory study. - The best current synthesis finds additive, not synergistic, effects of parasites on mortality, and no demonstrated interaction at colony level in the field. - The meta-analysis does support synergy among agrochemicals, which the chapter did not stress.

Implications for weight. The general point stands: multiple stressors matter, and single-stressor tests miss interactions. Environmental risk assessments that assume additivity may underestimate mixtures of agrochemicals [S43]. The specific Nosema claim should not be cited as established.

Claim 5: The precautionary bans worked: after France’s bans (sunflower 1999; maize and fipronil 2004), high summer mortality in intensive-farming areas stopped and colonies partly recovered; Italy’s 2008 moratorium cut reported poisonings from 185 to 2–3; Bayer disputes this, citing continued losses and no correlation in monitoring (pp. 386, 392–393, 401–402, 404; Table 16.1, p. 394)#

Original claim. - According to UNAF, summer mortality stopped and hives improved after 2003 (p. 386). The same page concedes that “no system exists in France for the accurate and extensive monitoring of honeybees”. - Table 16.1 reports cessation of summer mortality in 2005–2007, with “gradual” recovery and variable winter losses (p. 394). - Italy’s moratorium was followed by a fall in reported poisonings from 185 to two (p. 393) or three (p. 404). - Bayer cites large-scale monitoring that found “no correlation” (pp. 401–402).

Subsequent developments - No independent French evaluation found. I found no peer-reviewed or official evaluation, after 2013, of how the 1999/2004 French bans affected colony mortality. Fipronil was banned at the same time, which confounds attribution (digest caveat). - Losses remain multi-causal. - EPILOBEE (2012–2014, 17 countries, 5,798 apiaries) found winter losses of 2–32%. It identified beekeeper experience and disease control as major drivers; pesticide exposure was not a main focus [S65]. - Italy’s national monitoring network (2009–2010, after the moratorium) found average losses of 19% in 2009/10. In 2010 losses were related to pesticides found in bees, mostly organophosphates and pyrethroids in hive contents, and to the share of farmland nearby [S66]. - I did not locate an independent re-analysis of the Italian figure of 185 poisonings falling to two or three. The dust mechanism behind it is well confirmed (Claim 3). - Evidence on Bayer’s “no correlation”. - Against Bayer: across England and Wales (11 years), honeybee colony losses correlated with national imidacloprid use on oilseed rape [S40]. Wild-bee extinction rates in England rose with neonicotinoid use on oilseed rape [S39]. - For Bayer, as regards honeybee colonies: an independent Swedish landscape study found no detectable harm [S41], and a Bayer-funded German study found no effect [S42]. In a three-country trial, effects on honeybees were negative in Hungary and the UK and positive in Germany [S37]. - Exposure and wider recovery after bans. - Rural UK bumblebees had lower exposure after the moratorium [S56]. - French water hazard fell sharply after 2018 [S57]. - Insectivorous birds in France showed only “weak recovery” after the 2018 ban. The abundance gap between sites with high and no imidacloprid use narrowed only from 12.7% to 9% [S58]. - In Germany, applied toxicity to pollinators showed no trend, which “likely results from neonicotinoid use restrictions” offsetting other insecticides [S70].

Verdict: unclear. - The bans clearly reduced exposure. - Whether they restored honeybee colony health cannot be established from the evidence I found, because monitoring was weak and causes were confounded. - The chapter’s own admission that France lacked a monitoring system (p. 386) remains the key limitation. - Bayer’s “no correlation” claim was undermined at national scale in the UK, but it partly survives for honeybee colonies in field trials.

Implications for weight. - Treat “the bans worked” as plausible, not demonstrated, for honeybees. - The better-supported lesson is the chapter’s own: without a monitoring baseline, neither harm nor the effect of an intervention can be shown (digest insight 5). - Later bird data add a further point: removing a persistent agent does not bring quick biological recovery.

Claim 6: Banning seed-dressings did not measurably reduce French maize or sunflower yields (Fig. 16.3, p. 385); seed-dressing was used preventively even where pest risk was “low or zero” (pp. 383–384)#

Original claim. - Figure 16.3 (1995–2007) shows no fall in yield after the bans. 2007 was the best maize year, and the worst was the 2003 heatwave (p. 385). - The French oilseed technical institute (CETIOM) judged wireworm risk to sunflower “low or zero” for most French areas (p. 384). Use was preventive “regardless of the presence and abundance of pests” (p. 384). - The digest flags the “scale double standard” of using national yields.

Subsequent developments - Preventive use and small benefits. - A 2018 farm survey by the European Commission’s Joint Research Centre (JRC) covered eight EU case studies after the 2013 restrictions. It found that farmers [S67]: - switched to untreated seed in four cases and to unrestricted neonicotinoid or pyrethroid seed treatments in three; - increased soil or foliar sprays, mainly pyrethroids, in five; - perceived higher costs and pest pressure. - Swedish spring oilseed rape (23 trials): seed treatment reduced yield loss in one year of three [S68]. - US soybean (194 trials): “negligible” yield benefits [S69]. - English oilseed rape: seed coatings reduced later foliar sprays and “may derive an economic return” [S40]. - The sugar-beet counter-case. - France banned all neonicotinoids from 2018 under its 2016 biodiversity law. In 2020 aphid-borne virus yellows cut national beet yields from 851 q/ha (2019) to 649 q/ha, and output fell from 38 to 27 million tonnes. These are government figures citing the Agreste agricultural statistics service [S28, observations]. - A law of December 2020 on health emergencies in sugar beet allowed derogations for beet seed treatment until 1 July 2023. The Conseil constitutionnel upheld it (2020-809 DC, 10 December 2020) because it was limited to beet, to seed treatment and in time [S28]. - Derogations were granted for 2021 [S57] and, as I understand, for 2022 (not independently verified). - The CJEU (C-162/21, 19 January 2023) held that emergency authorisations under Article 53 cannot allow treated seed that an implementing regulation expressly bans [S18]. - Recurrence. - The 2025 “Duplomb” law’s broad reauthorisation of neonicotinoid-type substances was struck down (2025-891 DC, 7 August 2025). It was not limited in time, crop or method of application [S29]. - The 2026 emergency agriculture law allows the French agency for food, environmental and occupational health and safety (ANSES) to grant derogations, which the Conseil constitutionnel upheld with interpretive reservations on 14 August 2026 (2026-914 DC) [S30]: - flupyradifurone seed treatment for sugar beet; - flupyradifurone for cherries and apples, sprays included; - acetamiprid for hazelnuts. - The derogations last one year, renewable twice, with a three-year sunset. I have not verified whether any derogation has yet been granted. - These provisions entered the law through a Senate amendment of 3 July 2026, against organised opposition from beekeepers [S35]. - EU-wide pattern. - Member States granted 206 emergency authorisations for the three restricted substances in 2013–2019 [S1]. - EFSA found in 2017 that four of seven Member States examined had alternatives available or had not justified the danger [S1].

Verdict: partly held up. - Preventive overuse and weak average benefits have been confirmed in several crops. - The national-yield argument does not generalise. Where the seed treatment controlled a virus vector, the ban brought large losses (sugar beet, 2020). That drove six years of derogations and political reversal.

Implications for weight. Strong support for digest insight 12 (lock-in) and insight 15 (substitution by near-equivalents): pyrethroid sprays, unrestricted neonicotinoids, flupyradifurone and acetamiprid all filled the gap. Cost judgements need to be made crop by crop and pest by pest. Average national yields can hide concentrated losses, the mirror image of the chapter’s complaint about national-scale framing of bee losses.

Claim 7: Imidacloprid is “fit for the precautionary principle” given highly variable toxicity (21-fold oral LD50, 40-fold contact LD50, 1,000-fold chronic LOECs; half-lives from 83 days to 1–2 years; some effects “uncontrollable”); the honeybee is an “alarm bell” for other insect life (pp. 392–393)#

Original claim. - Effects are “highly variable” and depend on temperature, humidity and soil. “More than one study failed to find a dose-effect relationship”, so “some of the effects … are uncontrollable” (p. 392). - Honeybees are “unusually sensitive”, and their losses are “an ‘alarm bell’ of harm to other entomofauna” and to birds and plants (p. 393).

Subsequent developments - Variability and dependence on context: confirmed. - Toxicity is time-cumulative. Analysing the lethal dose against time for imidacloprid, a 2014 study found it scales with exposure time. Extrapolated to winter bees, 0.25 µg/kg in honey could be lethal to many bees near the end of life [S63] (abstract). - Effects varied by country [S37], by soil and rainfall [S53], and by temperature and Varroa load [S75]. - Sensitivity across bee species ranged widely (ratio 0.001 to 2,085.7), though about 95% of cases were within a factor of 10 of honeybees [S64]. - The 1,000-fold spread in chronic LOECs rests partly on Suchail (2001). The 2013 guidance used Suchail among the sources for its chronic test design [S1, Annex I]. I found no direct replication study. - Courts on precaution. The CJEU (2021) and the Conseil constitutionnel (2020, 2025, 2026) treated scientific uncertainty about neonicotinoids as grounds for restriction [S17, S28, S29, S30]. - The honeybee as sentinel: inverted at colony level. - A replicated Swedish landscape study: clothianidin-coated oilseed rape reduced wild-bee density, solitary-bee nesting and bumblebee colony growth, with no significant honeybee response. The authors warn that effects on honeybees “cannot always be extrapolated to wild bees” [S36]. - In the three-country trial, wild-bee reproduction was negatively correlated with residues, while honeybee effects varied by country [S37]. - A meta-analysis (2021) found field-realistic exposure reduced reproduction in non-Apis bees and bumblebee colony growth [S44]. - EFSA 2018 found low risk to honeybees for some uses but high or undemonstrated low risk once bumble and solitary bees were included [S13]. - Independent field work found honeybee colonies “relatively robust” [S41]. - Wider harm: confirmed. - Aquatic invertebrates: neonicotinoids are frequent in surface waters worldwide, with thresholds recommended below 0.035 µg/l for chronic exposure [S61]. This led to the EU water standards of 2026 [S24]. - Birds: Dutch insectivorous birds declined faster where imidacloprid levels in water were higher [S62]. French insectivorous birds were associated with imidacloprid use and recovered weakly after the ban [S58]. - An independent review of post-2013 evidence, by WIA-linked authors, reported harm to a wide range of non-target organisms from persistent low levels [S73] (abstract).

Verdict: partly held up. - “Fit for precaution”: strengthened. Variability, time-dependence and dependence on context are well documented, and courts have adopted precautionary reasoning. - The honeybee as a sensitive sentinel: weakened. At colony level honeybees are buffered and under-report harm, even though the harm to “other entomofauna” (and birds) that the chapter predicted was confirmed.

Implications for weight. The lesson on precaution holds. The sentinel lesson needs reversing: the managed, economically visible indicator species was among the least informative at population level. The choice of test organism shaped what counted as harm, a point that fits the chapter’s own argument about how framing and measurement shape evidence.

Original claim. - Original studies should be available “through a cost-free database” (p. 390). - Researchers’ conflicts of interest should be “readily available to the public”, and researchers should not assess products of their funders (p. 391). - Results should be published whatever they show, and early-warning scientists should be protected (p. 392).

Subsequent developments - Regulation (EU) 2019/1381 (Transparency and sustainability of EU risk assessment in the food chain) [S21]: - It applies to applications submitted from 27 March 2021. - Article 38(1) requires EFSA to publish proactively the “scientific data, studies and other information supporting applications”, subject to confidentiality carve-outs (Arts 39–39e). It also requires publication of experts’ annual declarations of interest. - Article 32b creates a database of studies commissioned by business operators. Applications relying on studies not notified in advance are invalid unless justified. This is the closest analogue to the chapter’s guarantee of publication, since it makes unfavourable studies harder to withhold. - Article 32c provides public consultation on the studies planned for renewals. Article 32d lets the Commission ask EFSA to commission verification studies “in exceptional circumstances of serious controversies”. - The public portal is OpenEFSA [S27]. - The recitals cite the glyphosate citizens’ initiative, not neonicotinoids, as the trigger [S21, recital 27]. - Whistleblowers and researchers. - Directive (EU) 2019/1937 protects people reporting breaches of EU law on environmental protection and on food and feed safety, among other areas. Transposition was due by 17 December 2021 [S22]. It protects reporting of breaches of law, not scientists pressured over their findings. - Directive (EU) 2024/1069 (anti-SLAPP, against strategic lawsuits against public participation) protects “researchers and academics”, among others, from abusive court proceedings. It covers only cross-border civil cases. Transposition was due by 7 May 2026 [S23, recital 6]. This bears directly on the defamation suits against beekeeper leaders (p. 380). - France’s Law 2013-316 of 16 April 2013 did three things [S31]: - gave a right to disclose in good faith serious risks to health or the environment; - created a national commission on professional ethics and alerts in public health and the environment; - required public, annually updated declarations of interest for that commission’s members. - Counter-current. The Commission’s December 2025 simplification proposal (COM(2025) 1030) would do the following [S26]: - make approvals of most active substances unlimited in duration (candidates for substitution excepted); - add targeted reassessment; - keep the ad hoc review under Article 21.

Periodic renewal is the channel through which new data on bees enter assessments, so this would weaken scheduled re-examination.

Verdict: held up. - The need for these measures was accepted, and most were enacted in EU and French law in 2013–2024. - Two recommendations were implemented only partly: publication guarantees (approximated by study notification) and protection of scientists specifically. - No evaluation yet shows whether the measures changed outcomes. - The main impetus came from other controversies.

Implications for weight. The recommendations are well aligned with where EU policy went, which strengthens their standing as general lessons. They should be presented as convergent with a broader reform movement rather than as effects of this case.

Claim 9: Bayer’s counter-claims (Panel 16.1, pp. 401–402): colony losses are multifactorial; large-scale monitoring found no correlation with neonicotinoid seed-dressings; Gaucho met all re-evaluation requirements; suspensions risk “stopping innovation” and should be reversible#

Original claim (Schmuck, pp. 401–402). - The authors took Gaucho “a priori” as “THE key cause”. - Monitoring programmes in Belgium, France, Germany and the US found no correlation. - Gaucho met all re-evaluation requirements, and “NO evidence of a causative link” was found. - Evidence from aphids showed bees would not be affected. - Suspensions risk “stopping innovation” unless reversible.

Subsequent developments - Multifactorial causes: held up. - EPILOBEE identified disease and beekeeping practice as drivers [S65]. Italy’s network found pathogens, pesticides and land use all mattered [S66]. - IPBES (2016) treats pollinator decline as multi-driver, while finding that neonicotinoids harm bees (as cited in [S19]; [S74]). - The chapter’s authors did not deny multiple causes. They argued multicausality was used to excuse inaction (p. 391). EU regulators acted on one tractable cause without resolving the others [S10, S14]. - No correlation: partly undermined. - Undermining it: the national-scale correlation in England and Wales [S40], wild-bee population effects [S39], negative honeybee effects in two of three countries in an industry co-funded trial [S37], and reduced honeybee health from chronic exposure near maize in Canada [S38]. - Consistent with it: independent and industry-funded studies of honeybee colonies [S41, S42]. - Met all requirements: overtaken. - The EU reviews (2013, 2018) found high risks under the updated method [S8, S12, S13]. - The approval lapsed in 2020 [S15]. - The courts rejected Bayer’s challenges [S16, S17]. - The requirements Gaucho met were those the auditors later called unsuitable for seed treatments [S1]. - Stopping innovation and reversibility: mixed. - EU restrictions were not reversed. - Newer systemic insecticides met similar concerns. Sulfoxaflor was restricted to permanent greenhouses in 2022 because risks to bumble and solitary bees from outdoor use could not be excluded [S20]. According to UNAF, France banned acetamiprid, sulfoxaflor and flupyradifurone by a decree of 16 December 2020, and the Conseil d’État rejected the industry challenge to that decree on 5 June 2025 [S34]. - Acetamiprid was renewed at EU level to 2033 on a finding of low risk to bees [S15]. So restrictions have been substance-specific rather than class-wide in the EU. - Reversibility did arrive politically, for near-equivalent substances. French derogations were granted in 2020–2022 and enacted again in 2026 [S28, S30]. The Commission’s 2025 proposal of unlimited approvals echoes the industry’s concern that repeated reassessment burdens innovation [S26]. - The aphid argument. Not tested in later literature that I found. The authors’ rebuttal (pp. 404–405) is consistent with the later finding that colony-level social buffering and species differences make extrapolation unreliable [S36, S49].

Verdict: weakened. - Bayer’s regulatory and evidential claims (“no evidence”, “met all requirements”) did not survive the EU reviews or the courts. - Two parts survive: the multifactorial framing, and the narrower claim that honeybee colonies often show no measurable harm in the field.

Implications for weight. The panel is best read as evidence of the dynamics the chapter describes. It shows how a party frames multicausality and cites the most favourable metric (honeybee colonies). The panel is not a reliable guide to the science. Its warning about innovation remains a live policy argument in 2025–2026.

Claim 10: A criminal investigation opened in Paris after UNAF’s 2001 complaint was still under investigation in March 2011 (p. 380)#

Original claim (p. 380). - The investigation “continues to stagnate”. Two judges had been replaced. - In March 2011 the judge was still deciding whether there would be a trial.

Subsequent developments (source: UNAF, a party. I could not access the court decisions themselves.) - Dismissal. An order of dismissal (non-lieu) was issued on 1 April 2014 after 13 years of investigation. According to UNAF, it was reasoned on “scientific uncertainty” about attributing increased bee mortality to Gaucho and about the scale of the mortality. UNAF appealed at once [S32]. - Appeals. - The Paris Court of Appeal confirmed the dismissal on 15 April 2015. The charges included deception, fraud and sale of toxic products. - The Cour de cassation rejected UNAF’s appeal on 4 January 2017. - UNAF counted five successive investigating judges over more than 16 years of proceedings [S33]. - UNAF’s framing. UNAF contrasts this with repeated annulments by the Conseil d’État of Gaucho authorisations between 1999 and 2006 [S33]. The chapter (p. 380) also records Conseil d’État findings against the Ministry.

Verdict: held up. The chapter’s description of a stagnating inquiry was accurate. The case ended with no trial.

Implications for weight. This is a useful addition to the chapter’s lessons. The standard of proof differs by forum: - Administrative and constitutional courts, and EU courts on the precautionary principle, accepted protective action under uncertainty [S17, S28]. - The criminal forum treated the same uncertainty as a bar to liability.

Accountability through the courts for early-warning harms of this kind is therefore weak. Most corrective action came through regulatory and administrative routes. It also supports digest insight 13: the costs of contesting evidence fell on the harmed party for more than 16 years.

Sources#

All retrieved 25 September 2026. “Abstract only” means I read the Europe PMC abstract, not the full text.

EU institutions: assessment method and regulatory decisions - [S1] European Court of Auditors. Special report 15/2020: Protection of wild pollinators in the EU – Commission initiatives have not borne fruit (July 2020; full text, including Box 6, paras 48–62, Annex I and the Commission replies). https://www.eca.europa.eu/Lists/ECADocuments/SR20_15/SR_Pollinators_EN.pdf - [S2] European Commission, DG SANTE. Protection of bees (web page on the bee guidance history: 2013 non-endorsement, the July 2019 vote, the October 2019 Parliament objection, the 2002 guidance “remains the basis”, the February 2021 protection-goal positions; content to May 2023). https://food.ec.europa.eu/plants/pesticides/protection-bees_en - [S3] EFSA PPR Panel. Scientific Opinion on the science behind the development of a risk assessment of plant protection products on bees. EFSA Journal 2012;10(5):2668 (23 May 2012; abstract only). https://doi.org/10.2903/j.efsa.2012.2668 - [S4] EFSA. Guidance on the risk assessment of plant protection products on bees (Apis mellifera, Bombus spp. and solitary bees). EFSA Journal 2013;11(7):3295 (4 July 2013; updated 4 July 2014; contents as summarised in [S1]). https://www.efsa.europa.eu/en/efsajournal/pub/3295 - [S5] EFSA. Revised guidance on the risk assessment of plant protection products on bees. EFSA Journal 2023;21(5):7989 (11 May 2023; abstract only), and EFSA news release “Bees and pesticides: updated guidance for assessing risks”, 11 May 2023. https://doi.org/10.2903/j.efsa.2023.7989 ; https://www.efsa.europa.eu/en/news/bees-and-pesticides-updated-guidance-assessing-risks - [S6] SCoPAFF, Section Phytopharmaceuticals–Legislation. Summary reports of 22 March 2023, 24 May 2023, 11 July 2023 and 29–30 June 2026, and the agenda for 24–25 September 2026. Index: https://food.ec.europa.eu/horizontal-topics/committees/paff-committees/phytopharmaceuticals_en. Reports: - 22 March 2023: https://food.ec.europa.eu/document/download/dca5c9f8-3432-4bf7-a14a-7000bb3faad3_en?filename=sc_phyto_20230322_ppl_sum.pdf - 24 May 2023: https://food.ec.europa.eu/document/download/6e46dd71-619d-4efe-ad5a-1520c3be1ceb_en?filename=sc_phyto_20230524_ppl_sum.pdf - 11 July 2023: https://food.ec.europa.eu/document/download/7098deed-80e4-4b6e-a40d-49945ec52d5a_en?filename=sc_phyto_20230711_ppl_sum.pdf - 29–30 June 2026: https://food.ec.europa.eu/document/download/e1f018bc-7c71-4a00-8880-bb58829b91ee_en?filename=sc_phyto_20260629_ppl_sum.pdf - Agenda, 24–25 September 2026: https://food.ec.europa.eu/document/download/05aa5e59-0506-4a3d-9745-6a09a5361b45_en?filename=sc_phyto_20260924_ppl_agenda.pdf - [S7] EFSA. Conclusions on the peer review of the pesticide risk assessment of the active substance proquinazid (rapporteur Sweden, co-rapporteur Lithuania). EFSA Journal 2026 (published 18 May 2026; full text via Europe PMC, PMC13181593). https://doi.org/10.2903/j.efsa.2026.10063 - [S8] EFSA. Press release on the risk assessment of three neonicotinoids for bees, 16 January 2013. https://www.efsa.europa.eu/en/press/news/130116 - [S9] EFSA. Conclusion on the peer review of the pesticide risk assessment for bees for the active substance imidacloprid. EFSA Journal 2013;11(1):3068 (abstract only). https://doi.org/10.2903/j.efsa.2013.3068 - [S10] Commission Implementing Regulation (EU) No 485/2013 of 24 May 2013 (clothianidin, thiamethoxam, imidacloprid; full text). https://eur-lex.europa.eu/eli/reg_impl/2013/485/oj - [S11] Commission Implementing Regulation (EU) No 781/2013 of 14 August 2013 (fipronil; full text). https://eur-lex.europa.eu/eli/reg_impl/2013/781/oj - [S12] EFSA. Press release on the updated neonicotinoid assessments for bees, 28 February 2018. https://www.efsa.europa.eu/en/press/news/180228 - [S13] EFSA. Peer review of the pesticide risk assessment for bees for the active substance imidacloprid considering the uses as seed treatments and granules. EFSA Journal 2018;16(2):5178 (February 2018; full text, PMC7009388). https://doi.org/10.2903/j.efsa.2018.5178 - [S14] Commission Implementing Regulation (EU) 2018/783 of 29 May 2018 (imidacloprid; permanent greenhouses only). https://eur-lex.europa.eu/eli/reg_impl/2018/783/oj - [S15] European Commission, DG SANTE. Neonicotinoids (approval history, including imidacloprid expiry on 1 December 2020 and acetamiprid renewal to 28 February 2033). https://food.ec.europa.eu/plants/pesticides/approval-active-substances-safeners-and-synergists/renewal-approval/neonicotinoids_en - [S16] General Court. Judgment of 17 May 2018, Bayer CropScience and Others v Commission, T-429/13 and T-451/13, EU:T:2018:280 (as summarised in [S17]). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62013TJ0429 - [S17] Court of Justice. Judgment of 6 May 2021, Bayer CropScience and Bayer v Commission, C-499/18 P (full text; paras 16, 20, 49, 56, 79–82). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62018CJ0499 - [S18] Court of Justice. Judgment of 19 January 2023, Pesticide Action Network Europe and Others, C-162/21 (full text). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62021CJ0162 - [S19] Commission Regulation (EU) 2023/334 of 2 February 2023 on maximum residue levels (MRLs) for clothianidin and thiamethoxam (recitals 7–11 cite IPBES 2016 and the WIA; applies from 7 March 2026). https://eur-lex.europa.eu/eli/reg/2023/334/oj - [S20] Commission Implementing Regulation (EU) 2022/686 of 28 April 2022 (sulfoxaflor restricted to permanent greenhouses). https://eur-lex.europa.eu/eli/reg_impl/2022/686/oj

EU governance, transparency and protection - [S21] Regulation (EU) 2019/1381 of 20 June 2019 on the transparency and sustainability of the EU risk assessment in the food chain (full text; Arts 32b–32d, 38, 39, recital 27; Art 10 on transitional measures). https://eur-lex.europa.eu/eli/reg/2019/1381/oj - [S22] Directive (EU) 2019/1937 of 23 October 2019 on the protection of persons who report breaches of Union law (Art 2, Art 26). https://eur-lex.europa.eu/eli/dir/2019/1937/oj - [S23] Directive (EU) 2024/1069 of 11 April 2024 on protecting persons who engage in public participation from manifestly unfounded claims or abusive court proceedings (recital 6, Art 22). https://eur-lex.europa.eu/eli/dir/2024/1069/oj - [S24] Directive (EU) 2026/805 of 30 March 2026 amending the Water Framework, Groundwater and Environmental Quality Standards Directives (OJ L, 20 April 2026; Annex, entries for acetamiprid, clothianidin, imidacloprid, thiacloprid and thiamethoxam; Art 4). https://eur-lex.europa.eu/eli/dir/2026/805/oj - [S25] European Commission. Impact assessment accompanying the proposal on surface-water and groundwater pollutants, SWD(2022) 540 final (26 October 2022; citing the EEA’s 2021 assessment of pesticides in water). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52022SC0540 - [S26] European Commission. Proposal for a Regulation on the simplification and strengthening of food and feed safety requirements, COM(2025) 1030 final, 16 December 2025 (explanatory memorandum and Art 14 as proposed). https://food.ec.europa.eu/document/download/b0817113-6edc-4219-b638-8060fee037d5_en?filename=horiz_omnibus_reg-com-2025-1030_en.pdf ; overview: https://food.ec.europa.eu/horizontal-topics/simplification-legislation_en - [S27] EFSA. OpenEFSA portal (landing page only). https://open.efsa.europa.eu/

France - [S28] Conseil constitutionnel. Décision n° 2020-809 DC du 10 décembre 2020 (sugar-beet derogation law; paras 18–24), and Government observations of 27 November 2020 (Agreste yield figures). https://www.conseil-constitutionnel.fr/decision/2020/2020809DC.htm ; https://www.conseil-constitutionnel.fr/sites/default/files/as/root/bank_mm/decisions/2020809dc/2020809dc_obs.pdf - [S29] Conseil constitutionnel. Décision n° 2025-891 DC du 7 août 2025, Loi visant à lever les contraintes à l’exercice du métier d’agriculteur (paras 75–83). https://www.conseil-constitutionnel.fr/decision/2025/2025891DC.htm - [S30] Conseil constitutionnel. Décision n° 2026-914 DC du 14 août 2026, Loi d’urgence pour la protection et la souveraineté agricoles (Art 6; paras 45–79). https://www.conseil-constitutionnel.fr/decision/2026/2026914DC.htm - [S31] Loi n° 2013-316 du 16 avril 2013 relative à l’indépendance de l’expertise en matière de santé et d’environnement et à la protection des lanceurs d’alerte (Légifrance; summary of Arts 1, 2, 6). https://www.legifrance.gouv.fr/loda/id/JORFTEXT000027324252 - [S32] UNAF. Un non-lieu dans l’affaire Gaucho. 22 April 2014 (party source). https://www.unaf-apiculture.info/un-non-lieu-dans-l-affaire-gaucho/ - [S33] UNAF. Communiqué de presse, GAUCHO illégal mais Bayer pas coupable !, 16 January 2017 (Cour de cassation, 4 January 2017; Cour d’appel de Paris, 15 April 2015; party source). https://www.unaf-apiculture.info/wp-content/uploads/2017/01/cp_2017-01-16_unaf_affaire_gaucho_bayer-118.pdf - [S34] UNAF. Le Conseil d’État confirme le bien-fondé scientifique de l’interdiction de l’acétamipride… 1 July 2025 (reporting Conseil d’État, 5 June 2025, rejecting the industry group Phyteis’s challenge to the decree of 16 December 2020; party source, court text not accessed). https://www.unaf-apiculture.info/le-conseil-detat-confirme-le-bienfonde-scientifique-de-linterdiction-de-lacetamipride-du-sulfoxaflor-et-du-flupyradifurone/ - [S35] UNAF. Vote du Sénat sur les néonicotinoïdes… (Senate vote of 3 July 2026 reintroducing acetamiprid and flupyradifurone), 6 July 2026, and L’UNAF mobilisée devant l’Assemblée nationale…, 21 July 2026 (party source; context for [S30]). https://www.unaf-apiculture.info/vote-du-senat-sur-les-neonicotinoides-lunaf-alerte-sur-un-passage-en-force-inacceptable-et-appelle-les-elus-a-rejeter-le-texte-pour-sauver-lapiculture-francaise-et-la-biodiversite/ ; https://www.unaf-apiculture.info/lunaf-mobilisee-devant-lassemblee-nationale-contre-le-retour-des-neonicotinoides/

Field studies of bees (abstracts via Europe PMC unless noted) - [S36] Rundlöf, M., et al. Seed coating with a neonicotinoid insecticide negatively affects wild bees. Nature 2015. https://doi.org/10.1038/nature14420 - [S37] Woodcock, B. A., et al. Country-specific effects of neonicotinoid pesticides on honey bees and wild bees. Science 2017 (funding: Bayer CropScience, Syngenta, NERC CEH, per Europe PMC grant metadata). https://doi.org/10.1126/science.aaa1190 - [S38] Tsvetkov, N., et al. Chronic exposure to neonicotinoids reduces honey bee health near corn crops. Science 2017. https://doi.org/10.1126/science.aam7470 - [S39] Woodcock, B. A., et al. Impacts of neonicotinoid use on long-term population changes in wild bees in England. Nature Communications 2016. https://doi.org/10.1038/ncomms12459 - [S40] Budge, G. E., et al. Evidence for pollinator cost and farming benefits of neonicotinoid seed coatings on oilseed rape. Scientific Reports 2015. https://doi.org/10.1038/srep12574 - [S41] Osterman, J., et al. Clothianidin seed-treatment has no detectable negative impact on honeybee colonies and their pathogens. Nature Communications 2019 (full text; publicly funded, no competing interests declared). https://doi.org/10.1038/s41467-019-08523-4 - [S42] Rolke, D., et al. Large-scale monitoring of effects of clothianidin-dressed oilseed rape seeds on pollinating insects in Northern Germany: effects on honey bees. Ecotoxicology 2016 (full text; “Funding of all expenses … through Bayer CropScience”). https://doi.org/10.1007/s10646-016-1725-8

Interactions and sublethal effects - [S43] Siviter, H., et al. Agrochemicals interact synergistically to increase bee mortality. Nature 2021. https://doi.org/10.1038/s41586-021-03787-7 - [S44] Siviter, H., Richman, S. K., Muth, F. Field-realistic neonicotinoid exposure has sub-lethal effects on non-Apis bees: a meta-analysis. Ecology Letters 2021. https://doi.org/10.1111/ele.13873 - [S45] Collison, E., Hird, H., Cresswell, J., Tyler, C. Interactive effects of pesticide exposure and pathogen infection on bee health – a critical analysis. Biological Reviews 2016. https://doi.org/10.1111/brv.12206 - [S46] Sánchez-Bayo, F., et al. Are bee diseases linked to pesticides? – A brief review. Environment International 2016. https://doi.org/10.1016/j.envint.2016.01.009 - [S47] Di Prisco, G., et al. Neonicotinoid clothianidin adversely affects insect immunity and promotes replication of a viral pathogen in honey bees. PNAS 2013. https://doi.org/10.1073/pnas.1314923110 - [S48] Grassl, J., et al. Synergistic effects of pathogen and pesticide exposure on honey bee (Apis mellifera) survival and immunity. Journal of Invertebrate Pathology 2018. https://doi.org/10.1016/j.jip.2018.10.005 - [S49] Odemer, R., Nilles, L., Linder, N., Rosenkranz, P. Sublethal effects of clothianidin and Nosema spp. on the longevity and foraging activity of free flying honey bees. Ecotoxicology 2018. https://doi.org/10.1007/s10646-018-1925-5 - [S50] Alonso-Prados, E., et al. Effects of thiamethoxam-dressed oilseed rape seeds and Nosema ceranae on colonies of Apis mellifera iberiensis under field conditions of central Spain. Insects 2022. https://doi.org/10.3390/insects13040371

Exposure, persistence and dust - [S51] Botías, C., et al. Neonicotinoid residues in wildflowers, a potential route of chronic exposure for bees. Environmental Science & Technology 2015. https://doi.org/10.1021/acs.est.5b03459 - [S52] Long, E. Y., Krupke, C. H. Non-cultivated plants present a season-long route of pesticide exposure for honey bees. Nature Communications 2016. https://doi.org/10.1038/ncomms11629 - [S53] Wintermantel, D., et al. Neonicotinoid-induced mortality risk for bees foraging on oilseed rape nectar persists despite EU moratorium. Science of the Total Environment 2020. https://doi.org/10.1016/j.scitotenv.2019.135400 - [S54] Humann-Guilleminot, S., et al. A nation-wide survey of neonicotinoid insecticides in agricultural land with implications for agri-environment schemes. Journal of Applied Ecology 2019. https://doi.org/10.1111/1365-2664.13392 - [S55] Mitchell, E. A. D., et al. A worldwide survey of neonicotinoids in honey. Science 2017. https://doi.org/10.1126/science.aan3684 - [S56] Nicholls, E., et al. Monitoring neonicotinoid exposure for bees in rural and peri-urban areas of the U.K. during the transition from pre- to post-moratorium. Environmental Science & Technology 2018. https://doi.org/10.1021/acs.est.7b06573 - [S57] Perrot, T., Bonmatin, J.-M., et al. Temporal and spatial trends of imidacloprid-related hazards in France. Science of the Total Environment 2024. https://doi.org/10.1016/j.scitotenv.2024.173950 - [S58] Perrot, T., et al. Weak recovery of insectivorous bird populations after ban of neonicotinoids in France, hinting at lasting impacts. Environmental Pollution 2025. https://doi.org/10.1016/j.envpol.2025.127132 - [S59] Pochi, D., et al. A device for pneumatic precision drills reducing the drift of the abrasion dust from dressed seed. Crop Protection 2015. https://doi.org/10.1016/j.cropro.2015.02.026 - [S60] Lin, C.-H., et al. Honey bees and neonicotinoid-treated corn seed: contamination, exposure, and effects. Environmental Toxicology and Chemistry 2021. https://doi.org/10.1002/etc.4957

Wider ecological effects and variability - [S61] Morrissey, C. A., et al. Neonicotinoid contamination of global surface waters and associated risk to aquatic invertebrates: a review. Environment International 2015. https://doi.org/10.1016/j.envint.2014.10.024 - [S62] Hallmann, C. A., et al. Declines in insectivorous birds are associated with high neonicotinoid concentrations. Nature 2014. https://doi.org/10.1038/nature13531 - [S63] Rondeau, G., et al. Delayed and time-cumulative toxicity of imidacloprid in bees, ants and termites. Scientific Reports 2014. https://doi.org/10.1038/srep05566 - [S64] Arena, M., Sgolastra, F. A meta-analysis comparing the sensitivity of bees to pesticides. Ecotoxicology 2014. https://doi.org/10.1007/s10646-014-1190-1

Colony monitoring - [S65] Jacques, A., et al. (EPILOBEE Consortium). A pan-European epidemiological study reveals honey bee colony survival depends on beekeeper education and disease control. PLoS ONE 2017. https://doi.org/10.1371/journal.pone.0172591 - [S66] Porrini, C., et al. The status of honey bee health in Italy: results from the nationwide bee monitoring network. PLoS ONE 2016. https://doi.org/10.1371/journal.pone.0155411

Farm benefits, alternatives and substitution - [S67] Kathage, J., et al. The impact of restrictions on neonicotinoid and fipronil insecticides on pest management in maize, oilseed rape and sunflower in eight European Union regions. Pest Management Science 2018 (European Commission JRC authors). https://doi.org/10.1002/ps.4715 - [S68] Lundin, O., Malsher, G., Högfeldt, C., Bommarco, R. Pest management and yield in spring oilseed rape without neonicotinoid seed treatments. Crop Protection 2020. https://doi.org/10.1016/j.cropro.2020.105261 - [S69] Mourtzinis, S., et al. Neonicotinoid seed treatments of soybean provide negligible benefits to US farmers. Scientific Reports 2019. https://doi.org/10.1038/s41598-019-47442-8 - [S70] Bub, S., et al. Trends of total applied pesticide toxicity in German agriculture. Environmental Science & Technology 2023. https://doi.org/10.1021/acs.est.2c07251

Context and authors’ network - [S71] Demortain, D. The science behind the ban: the outstanding impact of ecotoxicological research on the regulation of neonicotinoids. Current Opinion in Insect Science 2021. https://doi.org/10.1016/j.cois.2021.02.017 - [S72] van der Sluijs, J. P., et al. Conclusions of the Worldwide Integrated Assessment on the risks of neonicotinoids and fipronil to biodiversity and ecosystem functioning. Environmental Science and Pollution Research 2015 (co-authors include Bonmatin, Goulson, Krupke, Tapparo, Girolami). https://doi.org/10.1007/s11356-014-3229-5 - [S73] Wood, T. J., Goulson, D. The environmental risks of neonicotinoid pesticides: a review of the evidence post 2013. Environmental Science and Pollution Research 2017. https://doi.org/10.1007/s11356-017-9240-x - [S74] IPBES. Assessment Report on Pollinators, Pollination and Food Production (2016; landing page only; findings as cited in [S19]). https://www.ipbes.net/assessment-reports/pollinators - [S75] Monchanin, C., et al. Hazard of a neonicotinoid insecticide on the homing flight of the honeybee depends on climatic conditions and Varroa infestation. Chemosphere 2019. https://doi.org/10.1016/j.chemosphere.2019.02.129

Consulted, not relied on for specific claims - Veres, A., et al. An update of the WIA on systemic pesticides. Part 4: alternatives in major cropping systems. Environmental Science and Pollution Research 2020. https://doi.org/10.1007/s11356-020-09279-x - Furlan, L., et al. An update of the WIA on systemic insecticides. Part 3: alternatives to systemic insecticides. Environmental Science and Pollution Research 2021. https://doi.org/10.1007/s11356-017-1052-5