LL1-06 hindsight check: Ch6, PCBs and the precautionary principle#

Late lessons from early warnings: the precautionary principle 1896–2000 (EEA Environmental Issue Report No 22, 2001), chapter 6 by Janna G. Koppe and Jane Keys, pp. 64–75. Checked against developments up to 25 September 2026.

Overview#

The chapter’s structural argument has held up well after 25 years. In places it is stronger now than the authors could show at the time. Its prediction that “the toxic legacy will remain for many decades” (p. 69) has been borne out:

• the EU’s 2010 disposal deadline was still being worked through in many Member States in 2013–2015;
• on the Secretariat’s own 2023–2025 assessments, the Stockholm Convention’s 2025 target for removing PCB equipment from use was not on track, and in May 2025 the Parties were still being “urged” to meet it; donor-funded elimination programmes in several African countries were only starting up in early 2026;
• US rules still let PCB transformers run to the end of their lives;
• top marine predators still carry PCB loads linked to reproductive failure and projected population decline;
• in June 2026 EFSA cut its tolerable weekly intake for dioxins and dioxin-like PCBs again and found European dietary exposure above it in every age group.

The liability that Monsanto’s 1969 plan feared has also materialised. About a dozen US states, including Washington ($95 million, 2020), Oregon ($698 million, 2022) and Michigan and Rhode Island (at least $133 million combined, 2026), have settled with Monsanto over PCB contamination, without admission of liability (Claim 2).

The chapter says Monsanto held two positions at once: public doubt and private acknowledgement. The primary 1969 documents, released through litigation, confirm this.

Three parts of the chapter have aged less well:

• Its sourcing. Its best-known Monsanto “quotation” is partly a journalist’s paraphrase. Several dates are wrong: US and UK production end dates, and when Jensen identified PCBs. Its “100 years” framing starts the clock with a warning about a broader class of chemicals.
• Its health advocacy. It said that children’s behavioural and respiratory problems “could be due, in substantial part” to PCBs. Later pooled analyses, systematic reviews and WHO and EFSA assessments do not support this. They also do not support its attribution of neurodevelopmental effects to non-dioxin-like (“phenobarbital-like”) congeners, because those assessments stress that congener groups cannot be separated in human data. Later body-burden data contradict its rebuke of the Belgian toxicologists in 1999.
• Its 1930s counterfactual. The 1937 evidence was published in a journal, discussed with federal and state public-health officials, and read by the scientist who produced it as a workplace hazard that ventilation could control. Nothing known then pointed to persistence in the environment.

Net: the lessons about legacy stocks, the gap between the tested product and real exposure, and private knowledge versus public posture carry substantial weight. The graded-proof reading of history carries moderate weight. The specific paediatric attributions and the claim that 1930s precaution “would have prevented” the legacy carry little.

Method and limits#

• Primary documents. The 1969 Pollution Abatement Plan, the March 1969 Monsanto customer letter, the 1937 Drinker article with its published symposium discussion, and Monsanto’s 1979 history were read in the Toxic Docs litigation-document archive. EU legislation, Commission reports and Court of Justice (CJEU) judgments came from the EU Publications Office. I also used Stockholm Convention Secretariat documents; assessments by EFSA, WHO/JECFA, IARC and US EPA; US and Washington State court opinions; and peer-reviewed literature via Europe PMC and Crossref.
• Search tool unavailable for the first pass. The general web-search tool had used up its budget during the first pass, so sources were found through known repositories and bibliographic databases. A short second pass on 25 September 2026, with web search restored, added: the Oregon (2022) and Michigan/Rhode Island (2026) state settlements and the December 2025 settlement of Erickson (Claim 2); EFSA’s June 2026 update of its dioxin and dioxin-like PCB assessment (Claims 4 and 5); and the GEF-funded PCB elimination programme (Claim 7). Several company and state-government pages returned access errors, so some of these items rest on news reports or search-result extracts and are flagged as such.
• Annex 3 of the 2013 report was read from the EEA-hosted PDF.
• Abstract-only sources. Where I read only an abstract, the text says so. This applies mainly to Markowitz and Rosner (2018) and several epidemiological papers.

Claim 1. The two counterfactuals (p. 71)#

Original claim (p. 71).

• 1930s: there was already evidence, “some at a low level of proof”, that PCBs could poison people. It was “largely retained within the industry”, and applying the precautionary principle then “would have prevented the toxic legacy that now exists”.
• Late 1960s: the evidence was strong: Yusho, Jensen’s bioaccumulation findings, and the Baltic seals. Precautionary action at a level of proof below “beyond reasonable doubt” would “still have resulted in a more manageable, less costly problem”.

Developments since 2001.

The 1930s evidence was public, and it was framed as a workplace problem.

• The key 1930s evidence is the Drinker article and the discussion from the 30 June 1937 Harvard symposium, both published in the Journal of Industrial Hygiene and Toxicology in September 1937 (Drinker et al. 1937, via Toxic Docs).
• Those at the symposium included:
• R. R. Sayers (US Public Health Service, Division of Industrial Hygiene) and Louis Schwartz (USPHS dermatoses investigations);
• Connecticut and Massachusetts occupational-health officials;
• Monsanto’s R. Emmet Kelly.
• The participant list comes from a plaintiff-side key to the published record (Toxic Docs).
• Drinker found that chlorinated diphenyl was “probably the most dangerous” of the compounds he tested. He nonetheless concluded that these substances were “very little toxic” compared with benzene, and that the hazard could be controlled by ventilating to 0.5 mg/m³.
• By 1943 a New York State industrial-hygiene bulletin could list a substantial open literature on acne from chlorinated naphthalenes and diphenyls, including a 1936 American Journal of Public Health paper and Jones and Alden (1936) (Greenburg 1943, via Toxic Docs).
• So a precautionary response in 1937, by the standards of the time, would most plausibly have meant workplace controls, which is what Drinker proposed, not an end to production. Environmental persistence was not recognised until Jensen’s 1966 work.
• The chapter does not show how precaution about liver toxicity from mixed exposures in workers would have prevented a legacy that was environmental.

The late 1960s option to stop was real, and it was explicitly weighed and rejected.

• Monsanto’s November 1969 “PCB Environmental Pollution Abatement Plan” set out “Discontinue Manufacture” as one option and rejected it (Toxic Docs; rough draft dated 10 November 1969: Toxic Docs). See Claim 2.
• Fast partial action was clearly feasible at the 1969 level of proof. Monsanto’s own 1979 account says it dropped plasticiser, hydraulic and heat-transfer uses in 1970–72. It gives a 64% fall in its PCB sales between 1970 and 1972 (Monsanto, 16 November 1979, via Toxic Docs).
• Dielectric (closed) uses were a different story:
• They continued until 1977 in the US.
• EPA’s 1979 rule let existing electrical equipment holding about 578 million of the 750 million pounds of PCBs then in use continue in service for its lifetime (EPA press release, 19 April 1979).
• Decades later, the largest reported source of remaining EU PCB air emissions (52%) was still the use and disposal of PCB electrical equipment (Commission Union Synthesis Report SWD(2022)291, 16 September 2022).
• This supports the chapter’s 1960s counterfactual. Much of the later legacy traces to closed uses whose phase-out was deferred.

A trade-off the chapter omits.

• The 1969 plan framed PCB use in terms of fire safety. It posed a choice between risking human lives through fire and risking the extinction of some bird species.
• The US government’s 1972 Interdepartmental Task Force judged continued use in transformers and capacitors “necessary” because of fire and explosion risks. This is as quoted in Monsanto’s 1979 account; I did not see the original report.
• Monsanto’s account also says that silicone fluids for transformers and phthalate esters for capacitors were adopted around 1976 once substitution was pushed. So the alternatives existed within a few years, which is consistent with the chapter’s remark on p. 72.

Later scholarship.

• Markowitz and Rosner, historians who have testified for plaintiffs, reviewed the litigation archive. They argue that Monsanto understood the “world-wide ecological problem” and kept producing (J Public Health Policy 39(4):463–540, November 2018; abstract only read).

The EEA’s own update.

• Koppe’s 2013 update in Annex 3 of Late Lessons II (p. 727) does not revisit the counterfactual. It adds claims about diabetes, obesity and epigenetics, and says that “denial” continues (EEA Report 1/2013, annexes, archived PDF).

Verdict: partly held up.

• 1930s counterfactual: weakened. The evidence was occupational, confounded by mixed exposures, openly published, and answered with ventilation standards.
• Late-1960s counterfactual: plausible and supported by the primary record, but unquantified. It passes over a real risk–risk trade-off.

Implication for weight. “Acting at below-conclusive proof in the late 1960s would have shrunk the legacy” can carry moderate weight. “Precaution in the 1930s would have prevented it” should carry little.

The more defensible 1930s lesson is about framing. A hazard that was defined as occupational and controllable was never examined for its environmental fate (pp. 64, 71). The problem was not secrecy.

Claim 2. Monsanto’s 1969 “Pollution Abatement Plan” (p. 65)#

Original claim (p. 65). Publicly, Monsanto called toxicity claims “simply not true”. Privately, its 1969 plan admitted contamination across North America, Europe and remote regions. It rejected stopping production because this would cause “profits to cease and liability to soar because we would be admitting guilt by our actions”. The chapter’s only source is Francis (1998).

Developments since 2001.

The primary document can now be read.

• The plan survives in several copies in the litigation record. One version begins with a 15 October 1969 Ad Hoc Committee report (committee members Farrar, Hodges, John, Richard and Wheeler). A rough draft is dated 11/10/69, that is, 10 November 1969 (Toxic Docs kZv8Ew16…; Toxic Docs 5LOjyE97…).
• The Bates numbers include “NEV 021436–021456”, which suggests this is the Nevada Power litigation exhibit that Francis reproduced as “Exhibit 422”.

What the document confirms.

• The “Extensiveness” passage is quoted accurately. It covers the entire United States, Canada, the UK and Sweden, then Europe, Asia and Latin America, and says contamination has been shown in “some of the very remote parts of the world”. The only difference is minor: the original reads “have been shown”.
• Under option B, “Discontinue Manufacture of All Polychlorinated Biphenyls”, the plan says, word for word: “We would be admitting guilt by our actions.”
• The public position is confirmed from a primary document. The customer letter of 3 March 1969, signed by Elmer P. Wheeler, Manager of Environmental Health, says the “highly toxic” charge “is simply not true”. It questions whether the residues are PCBs at all, and calls for “extensive research, on a worldwide basis” (Toxic Docs).
• The same person was on both sides. “E. Wheeler” also sat on the October 1969 Ad Hoc Committee. That report fed a plan which states that the evidence of persistence and universal presence “is beyond questioning”. This strengthens the chapter’s “two positions” point.

What the document does not contain.

• The phrase “profits to cease and liability to soar” is not in the plan’s text.
• Francis’s article is a web reprint of his Sierra cover story (September/October 1994). In it, only “we would be admitting guilt by our actions” sits inside quotation marks. The rest is Francis’s description of a page of “Profit and liability vs. time” charts (Wayback capture, 5 December 2000).
• In both copies examined, those chart axes are blank. The EEA chapter merged paraphrase and quotation into a single quoted phrase.

Context the chapter leaves out.

• The plan gives several reasons for rejecting discontinuation:
• the business would be lost, with few substitutes;
• contract and capital losses;
• competitors would gain;
• liability would remain anyway, because “the pollutants are present in the environment”;
• moving to lower-chlorinated products might already cut contamination to “an acceptable level”.
• The plan also:
• says discontinuation “could be an eventuality”;
• recommends phasing out the higher-chlorinated PCBs in many applications;
• frames the fire-safety trade-off;
• says there was “no harmful effect known to man” after 40 years, written the year after the Yusho poisoning;
• asks the company to prepare for a forced sudden withdrawal (a “pullout”) or drastic legislative action, while judging these unlikely.

Litigation since 2001.

• Washington State sued in 2016, citing the 1969 committee documents. Monsanto settled for $95 million on 24 June 2020. The same release reports a separate $550 million class-settlement pool for 2,528 governmental entities, subject to court approval (Washington AG release, 24 June 2020).
• Oregon settled for $698 million on 15 December 2022, the state’s largest environmental damage recovery. The state’s complaint (filed 2018) alleged that Monsanto knew PCBs were harmful “since at least 1937”. Monsanto admitted no wrongdoing and said it had “voluntarily ceased production of PCBs in 1977” (OPB, 15 December 2022; secondary, because the Oregon Department of Justice release could not be retrieved).
• Michigan and Rhode Island settled on 18 May 2026 for a combined minimum of $133 million: at least $108 million for Michigan, which could rise to $240 million, and at least $25 million for Rhode Island. The settlements were made “without any admission of liability or wrongdoing”. Reports put the number of US states that have settled with Monsanto over PCBs at twelve, including Illinois and West Virginia in December 2025 (Michigan AG release, 18 May 2026, not retrievable directly; details from Great Lakes Now, 20 May 2026, secondary).
• Sky Valley teachers. The Washington Supreme Court reinstated a jury verdict of about $185 million ($50.15 million compensatory, $135 million punitive) for three teachers. The case was against Pharmacia LLC, which the court describes as successor-in-interest to the original Monsanto Company. The opinion recites that Monsanto’s own 1930s testing indicated risks of systemic toxicity (Erickson v. Pharmacia LLC, No. 103135-1, 30 October 2025, amended 21 January 2026).
• Bayer, Monsanto’s parent, reports that Monsanto settled Erickson on confidential terms on 10 December 2025. In August 2025 it had announced agreements in principle to resolve the other Sky Valley cases, more than 200 plaintiffs in all (Bayer, “Managing and Mitigating the U.S. PCB Litigation Risk”; the page returned an access error, so the wording is taken from search-result extracts of it).
• The pattern. Since 2020 the large state and municipal cases reviewed here have ended in settlements without admission of liability. The historical documents have shaped pleadings, trials and settlement values. Among the cases reviewed here, Erickson is the one where a jury verdict that included punitive damages was upheld on appeal, and it then settled. Jury verdicts in the other Sky Valley trials were not checked in this pass.
• For balance: some juries found for Monsanto. A Los Angeles jury did so in the first PCB–lymphoma trial (Law360, 22 May 2014, secondary, archived), as did a St Louis County jury in a multi-plaintiff cancer trial (St Louis Post-Dispatch, 7 July 2015, secondary, archived). These cases turned on whether PCBs caused the plaintiffs’ cancers, not on what the company knew.

Verdict: partly held up.

• The substance is confirmed from the primary record and has been central to later litigation.
• The quoted wording is inaccurate.
• The document’s reasoning is more mixed than the chapter suggests: several commercial and legal reasons, a fire-safety framing, and a recommendation for partial phase-out.

Implication for weight. The mechanism is:

• a firm acknowledges a problem privately while calling publicly for more research;
• liability makes any visible admission, including withdrawing the product, costly.

This mechanism can now carry strong weight, because it rests on the primary document rather than a secondary article (p. 65). Any reuse should quote the plan itself and drop the “profits to cease and liability to soar” wording.

Claim 3. Late-1930s awareness, and the “mob hysteria” remark (p. 64)#

Original claim (p. 64). By the late 1930s Monsanto “was certainly aware of adverse health effects in workers exposed to PCBs”. The evidence given is the 1936 Halowax cases and the 1937 Drinker meeting with Monsanto, General Electric and the US Public Health Service. Halowax’s president, Sanford Brown, closed the meeting by stressing the “necessity of not creating mob hysteria”, as quoted via Francis.

Developments since 2001.

Earlier awareness than the chapter states.

• Monsanto’s own 1979 review says that Swann Chemical saw dermatological symptoms in workers exposed to chlorinated biphenyls “sometime before 1934”. In May 1934 Swann asked F. B. Flinn of Columbia to investigate (Monsanto 1979, via Toxic Docs).
• Monsanto bought Swann in 1935 (Monsanto 1979; Erickson, 2025).
• Jones and Alden’s “An acneform dermatergosis” appeared in the Archives of Dermatology and Syphilology in June 1936 (doi:10.1001/archderm.1936.01470120073010; bibliographic record only). Monsanto’s 1979 review says the Swann studies were published in 1936.

The 1937 record.

• The Drinker article and symposium discussion are primary sources (Toxic Docs).
• Rats exposed to chlorinated diphenyl developed liver lesions.
• Monsanto’s Kelly told the meeting that Monsanto plants had seen skin eruptions of uncertain cause, but no systemic reactions in its workers.

“Mob hysteria” is verified word for word, and it was published.

• In the published symposium discussion, Brown speaks of “the necessity of not creating mob hysteria on the part of the workmen in the plants where these inspections are made”. The context is state inspections.
• In the same remarks he refers to a “comprehensive program” being planned with Monsanto.
• So the chapter’s quotation is accurate. But the remark appeared in a journal, not in a private record.

The deaths were mostly not PCB-specific.

• Drinker’s case descriptions show that only one of the three fatal Halowax cases (Patient 1) had been exposed to a mixture containing chlorinated diphenyl, at about 10%. The other two had been exposed to penta- and hexachloronaphthalenes.
• Monsanto’s 1979 review likewise notes that the 1930s–40s literature mostly concerned mixtures with chlorinated naphthalene.

Verdict: held up. Awareness of worker harm is documented, earlier than the chapter says, and the quotation is accurate. There are two caveats: the fatal cases were mostly not PCB-specific, and the “hysteria” remark was published rather than private. Both weaken the chapter’s later inference that 1930s evidence was “largely retained within the industry” (p. 71).

Implication for weight. “Producers knew of worker harm by the mid-1930s” is strong. “Warnings stayed inside a closed circle” is weak for the 1930s. The evidence was published, and federal and state public-health officials were in the room. What went wrong was how the hazard was framed and followed up.

Claim 4. Neurodevelopmental effects at background exposure, attributed to non-dioxin-like congeners (pp. 69–70, Box 6.3)#

Original claim (pp. 69–70).

• Prenatal exposure to background levels of PCBs in the US, Canada and western Europe causes adverse neurodevelopmental effects.
• These effects are “related to … phenobarbital-like PCBs and not to dioxin-like PCBs”.

Developments since 2001.

Support for an association.

• A review of nine prospective birth cohorts found the most consistent effect to be impaired executive function (Boucher et al., EHP 117:7–16, January 2009).
• In the Michigan cohort, deficits in attention and working memory at age 11 were seen mainly in children who had not been breastfed (Jacobson & Jacobson, J Pediatr 143:780–788, December 2003).
• A 2024 systematic review of 87 studies found evidence for poorer cognition and attention in middle childhood. It found no or negligible links with early motor development or with ADHD and autism, and marked heterogeneity between studies (Balalian et al., Environ Res 252:118912, July 2024; abstract read).

Evidence against, or inconsistent.

• The US Collaborative Perinatal Project (births 1959–65, about 900 mother–child pairs, relatively high exposure) found no association with IQ at age 7 (Gray et al., Am J Epidemiol 162:17–26, July 2005).
• A pooled analysis of seven European cohorts (4,437 pairs) found no association between PCB-153 and ADHD (Forns et al., Int J Epidemiol 47:1082–1097, August 2018).
• The Dutch cohort in Box 6.3, followed up:
• At age 9, higher prenatal PCB exposure lengthened P300 latencies, a brain-response measure, while longer breastfeeding shortened them (Vreugdenhil et al., Dev Med Child Neurol 46:398–405, June 2004).
• At ages 13–15, most exposures had no clinically relevant consequences (Berghuis et al., Environ Int 121:13–22, December 2018).
• A 2003 review already noted that the major congeners are so highly correlated with each other and with total PCBs that their separate roles are hard to separate statistically (Schantz et al., EHP 111:357–576, March 2003).

What WHO and EFSA concluded.

• WHO’s Joint FAO/WHO Expert Committee on Food Additives (JECFA), 2016:
• Some well-conducted studies suggest potential neurodevelopmental effects linked to non-dioxin-like PCBs.
• Because non-dioxin-like and dioxin-like exposures are strongly collinear, their independent effect is “very difficult” to estimate.
• Studies whose PCB sums include PCB 118, as the Dutch sum in Box 6.3 does, “do not provide any relevant information” on effects specific to non-dioxin-like PCBs.
• In rodents, neurodevelopment is not the most sensitive endpoint.
• Margins of exposure for adults and children were judged adequate. For breastfed infants, the benefits of breastfeeding “are considered to outweigh” possible disadvantages (WHO Food Additives Series 71-S1, issued 15 July 2016).
• EFSA, 2018, on dioxins and dioxin-like PCBs:
• The neurodevelopmental evidence was “not sufficient to form a basis for the risk assessment”.
• The critical effect became semen quality.
• The tolerable weekly intake (TWI) was cut from 14 to 2 pg TEQ/kg body weight per week, where TEQ means dioxin toxic equivalents.
• European exposures considerably exceed this new TWI (EFSA Journal 16(11):5333, November 2018).
• WHO, 2022, and EFSA, 2026, on dioxin-like potency:
• A WHO expert panel re-evaluated the toxic equivalency factors (TEFs) for dioxins and dioxin-like PCBs in October 2022, this time using a relative-potency database and Bayesian modelling rather than expert consensus alone (DeVito et al., Regul Toxicol Pharmacol 146:105525, January 2024).
• Using the new TEFs, EFSA set a TWI of 0.6 pg TEQ/kg body weight per week on 10 June 2026. The critical effect is again developmental male reproductive toxicity, derived from a rat study with TCDD, the most toxic dioxin. Neurodevelopment is not the critical endpoint.
• The new TEFs lower estimated European exposure by about 27–35%, but exposure still exceeds the TWI in every age group, most of all in toddlers and children. Levels in human milk also exceed it (EFSA plain-language summary, 10 June 2026; EFSA news, 10 June 2026; full opinion doi:10.2903/j.efsa.2026.10103, not read).
• EFSA’s assessment covers only the dioxin-like group, so it does not test the chapter’s congener attribution directly. It does show that European regulatory concern about background exposure now centres on dioxin-like developmental effects on male reproduction, not on neurodevelopment.
• WHO/UNEP breast-milk surveys:
• Levels are still above toxicologically “safe” values.
• In-utero exposure matters more than lactational exposure.
• The advantages of breastfeeding “far outweigh” possible disadvantages (van den Berg et al., Arch Toxicol 91:83–96, January 2017).

Verdict: partly held up.

• Associations at background exposure have partial, domain-specific support, in cognition and attention in middle childhood.
• They are inconsistent, small and partly transient, and neither WHO nor EFSA uses them to set limits.
• The attribution to non-dioxin-like (“phenobarbital-like”) congeners is not supported. Later assessments say the data cannot separate the congener classes.
• The chapter’s caution about breastfeeding advice (p. 69) matches later WHO and JECFA judgements.

Implication for weight.

• Moderate for “prenatal exposure at background levels may subtly affect some cognitive functions”.
• Low for the mechanism.
• The episode itself carries a lesson. The chapter warns that lumping a heterogeneous class together produces confusion (pp. 66–68). It then over-reads mechanism from correlated exposures, which is a mirror-image error.

Claim 5. Behavioural and respiratory problems “in substantial part” due to PCBs, and an “urgent need” to reduce body burdens (p. 72)#

Original claim (p. 72). Behavioural problems and respiratory diseases in children “could be due, in substantial part, to intoxication with PCBs”. There is “an urgent need to find ways of reducing current body burdens”.

Developments since 2001.

Behaviour.

• The pooled European analysis found no ADHD association for PCB-153 (Forns et al. 2018).
• The 2024 systematic review found little or no link with ADHD or autism, though it did find attention effects in middle childhood (Balalian et al. 2024).
• One autism meta-analysis reports positive associations for some congeners, but grades the certainty as low to moderate (Xu et al., Environ Pollut 336:122439, November 2023).

Respiratory disease.

• A systematic review of 41 studies found “limited” evidence linking prenatal PCBs to respiratory infections. It found no association between postnatal PCBs and asthma-related symptoms (Gascon et al., Environ Int 52:51–65, February 2013).
• A meta-analysis of ten European cohorts (4,608 children) found PCB-153 associated with bronchitis before 18 months (relative risk 1.06 per doubling of exposure). The association weakened when DDE, a DDT breakdown product, was included in the analysis, and there was no association with wheeze (Gascon et al., Epidemiology 25:544–553, July 2014).
• Clearer immune signals come from high-exposure populations:
• reduced antibody responses to vaccination in Faroese children, whose diet includes whale blubber (Heilmann et al., PLoS Med 3:e311, August 2006);
• more ear and lower-respiratory infections among the most exposed Inuit children (Dallaire et al., EHP 114:1301–1305, August 2006).
• JECFA (2016) notes associations with acute respiratory infections in some cohorts.
• No authoritative assessment reviewed here (JECFA 2016, EFSA 2018 and 2026, Gascon 2013) attributes a substantial share of childhood behavioural disorders or respiratory disease to PCBs.

Body burdens.

• Body burdens have fallen steeply. In Stockholm mothers’ milk, TEQ from dioxins and dioxin-like PCBs fell by 5.8–6.8% a year between 1972 and 2011 (Fång et al., Environ Int 60:224–231, October 2013). WHO/UNEP surveys show downward trends.
• The reductions came from source controls and time, not from interventions on people’s existing body burdens.
• The chapter’s underlying worry about low-level involuntary exposure is echoed in:
• EFSA’s 2018 finding that European exposure still considerably exceeds the new TWI (EFSA 2018);
• EFSA’s June 2026 update, which cut the TWI again (to 0.6 pg TEQ/kg body weight per week on the 2022 TEFs), found exposure above it in every age group and in human milk, and flagged women of childbearing age because of possible effects on male offspring exposed via the mother (EFSA news, 10 June 2026);
• WHO/UNEP’s 2017 call for further measures directed at sources (van den Berg et al. 2017).

Verdict: weakened. The attribution of a “substantial part” of these conditions to PCBs has not been borne out. The broader concern about reducing exposure has been vindicated in a different form, through successively lower tolerable intakes (most recently EFSA, June 2026) and source control, with the concern now centred on reproductive rather than behavioural or respiratory effects.

Implication for weight. Low as an empirical attribution. Moderate as a normative point that involuntary prenatal exposure deserves protective policy even when effects are uncertain (p. 72).

Claim 6. Bioaccumulated residues are more toxic than commercial mixtures (p. 67)#

Original claim (p. 67). Bioaccumulated PCBs “appear to be more toxic than commercial PCBs”. Mink fed Great Lakes fish showed liver and reproductive toxicity comparable to three times the dose of Aroclor 1254 (Hornshaw et al. 1983).

Developments since 2001.

EPA adopted the claim.

• The US EPA IRIS cancer assessment for PCBs was last revised on 1 October 1996 and is still the current IRIS entry. It says that bioaccumulated PCBs “appear to be more toxic than commercial PCBs”, citing the same two papers as the chapter (EPA IRIS summary).
• It puts food-chain exposure in its highest “high risk and persistence” tier. The upper-bound slope factor is 2.0 per mg/kg-day, against 0.4 and 0.07 for the lower tiers.

Mink studies replicated the pattern.

• Mink fed Saginaw Bay carp had impaired reproduction and kit survival at 0.72 mg total PCB per kg of diet. The lowest observed adverse effect level was 0.134 mg/kg body weight per day (Heaton et al., Arch Environ Contam Toxicol 28(3), April 1995).
• Housatonic River fish at about 1% of the diet were predicted to cut kit survival by 20% or more (Bursian et al., Environ Toxicol Chem 25:1533–1540, June 2006).
• A quantitative analysis of more than 50 mink tests found that internal, bioaccumulated dose explained reproductive outcomes better than dioxin-like TEQ (Fuchsman et al., Arch Environ Contam Toxicol 54:130–144, January 2008).

Caveat. Fish from contaminated waters carry co-contaminants such as dioxins, furans, mercury and DDE. Comparisons of potency depend on the dose metric chosen.

Verdict: held up, and taken up in regulatory practice.

Implication for weight. Strong for the lesson that the tested commercial product may not represent the real exposure once the environment has transformed it (p. 67).

Claim 7. The legacy will last decades; transformers still in use; the EU 2010 phase-out (pp. 69, 72)#

Original claims (pp. 69, 72).

• “The toxic legacy will remain for many decades.”
• A large share of historic production has “escaped beyond our control”.
• PCB transformers still in use will leak as they age.
• Directive 96/59/EC requires phase-out by 2010.

Developments since 2001.

EU compliance.

• The deadline. Directive 96/59/EC, Article 3, requires inventoried equipment (more than 5 dm³ of PCBs) to be decontaminated or disposed of “at the latest by the end of 2010” (Directive 96/59/EC, 16 September 1996).
• Early enforcement. The Court of Justice ruled against Member States for failing their inventory or planning obligations:
• Italy (C-46/01, 27 February 2002);
• France (C-177/01, 6 June 2002);
• Spain (C-47/01, 3 October 2002);
• Luxembourg (C-174/01, 5 December 2002);
• Greece (C-83/02, 5 June 2003);
• Malta (C-508/06, 29 November 2007).
• Verification. In October 2010 the Commission said it would verify implementation after the deadline (COM(2010)562, 15 October 2010).
• The Commission’s 2022 synthesis report (SWD(2022)291, 16 September 2022):
• For 2013–2015, disposal of the larger equipment “was still ongoing in many cases”.
• Member States estimated how much of their 1990 PCB volume was still in use in 2015:
  • under 1% in Belgium (Flanders), Denmark, Germany, Finland and Sweden;
  • about 3% in Lithuania and Portugal;
  • 6–10% in Czechia;
  • 7–8% in the UK and the Netherlands;
  • 19% in Ireland;
  • 30% in Croatia (against a 2008 baseline);
  • 49% in Romania (against a 2005 baseline).
• Czechia reported 851 PCB transformers still in place in September 2016.
• National plans also flag PCBs in buildings: 17–87 t in Denmark in 2013, 20–50 t in Sweden, and an estimated 400 t in Lithuanian buildings from 1950–70.
• EU PCB air emissions fell by about 80% between 1990 and 2015. The largest remaining source (52%) is the use and disposal of PCB electrical equipment.
• A new EU deadline. The EU POPs Regulation now requires Member States to identify and remove from use equipment containing more than 0.005% PCBs and more than 0.05 dm³ “no later than 31 December 2025” (Regulation (EU) 2019/1021, 20 June 2019, Annex I).

Global progress under the Stockholm Convention.

• Targets. Parties must eliminate PCB use in equipment by 2025 and manage PCB liquids and equipment in an environmentally sound way by 2028 (Stockholm Convention PCB overview).
• The Secretariat’s 2023 progress report (UNEP/POPS/COP.11/INF/11, 18 April 2023):

• The waste figure is the weight of equipment and contaminated material, not pure PCB.
• 107 Parties had not submitted their fifth national report, and 64 had never reported.
• The report says elimination must be accelerated to meet the 2025 and 2028 goals.
• The revised 2025 strategy repeats that “elimination of PCBs needs to be accelerated” (UNEP/POPS/COP.12/INF/11, 3 February 2025).
• COP-12, May 2025. Decision SC-12/3 again “urges” Parties to meet the 2025 and 2028 goals. The next formal review is scheduled for COP-13 (SC-12/3).
• Nothing reviewed indicates that the 2025 goal was met globally. The Secretariat’s own assessments indicated it was off track, although no formal post-deadline review has yet been published. In September 2026 the Secretariat’s progress page still showed only a 2020 global snapshot and pointed to a new review at COP-13 (Progress towards PCB elimination, consulted 25 September 2026).
• Donor-funded work towards the 2025 and 2028 goals was still starting up in 2026. A GEF-funded Global Elimination Program for PCBs, run with the World Bank, African Development Bank, UNDP and UNEP, aims to eliminate 8,750 tons of PCBs (the article does not say whether this is pure PCB or contaminated oil and equipment) in Cameroon, Eswatini, Gabon, Madagascar, Nigeria and Uganda by replacing contaminated utility equipment. GEF reported data quality, tracking and disposal capacity as obstacles (GEF feature, 24 April 2025). Its inception workshop was held in Abidjan on 27–29 January 2026, where participants reported capacity and financing gaps (African Development Bank news item, not retrievable directly; details from search-result extracts).

The United States.

• The US signed the Stockholm Convention on 23 May 2001 and has not ratified it (UN Treaty Collection).
• Federal rules still allow PCBs “at any concentration” in transformers for “the remainder of their useful lives”. Some restrictions apply; for example, certain network transformers in or near commercial buildings have been banned since 1 October 1990 (40 CFR 761.30, eCFR as of 1 September 2026).

Environmental trends.

• The Global Monitoring Plan, as summarised by the Secretariat in 2025, says concentrations are declining and “starting to level off”. Exceedances remain:
• PCB-118 close to or above the assessment criterion in three of five OSPAR (north-east Atlantic) assessment areas;
• parts of the Baltic;
• some Great Lakes fish;
• Mediterranean sperm whales, which carry the highest dioxin-like PCB levels recorded for the species;
• Arctic top predators, whose levels are still high and “associated with negative effects” (UNEP/POPS/COP.12/INF/11).
• PCBs in European cetaceans declined after the mid-1980s EU ban but have since stabilised. Mean levels in striped dolphins, bottlenose dolphins and killer whales exceed known toxicity thresholds (Jepson et al., Sci Rep 6:18573, 14 January 2016).
• PCB effects are projected to threaten the long-term viability of more than half the world’s killer whale populations (Desforges et al., Science 361:1373–1376, 28 September 2018).

Recovery after restriction, which Figure 6.1 shows but the text does not discuss.

• Baltic grey seals. Between 1977–86 and 1987–96, uterine obstructions fell from 42% to 11% and pregnancy rates rose from 9% to 60% (Bergman, APMIS 107:270–282, 1999).
• Swedish white-tailed eagles. Breeding productivity recovered once PCB and DDE levels in eggs fell below about 800 and 300 µg/g lipid. PCBs still averaged 250–500 µg/g in 1996–2005 (Helander et al., Ambio 37:425–431, September 2008).

How much “escaped”.

• In 1979 EPA estimated that, in the US alone, 150 million pounds of PCBs were dispersed in the environment and 290 million pounds were in landfills, against 750 million pounds still in use (EPA, 19 April 1979).
• The global Stockholm totals are consistent with a large share of production remaining unaccounted for. The Secretariat warns that national inventories are inaccurate, so no precise fraction can be given.

The EEA’s own update. Annex 3 of the 2013 report (p. 727) does not address transformers or deadlines.

Verdict: strengthened. Every element of the legacy prediction has been borne out. Deadlines slipped, top predators remain affected, and the installed equipment and buildings remain the dominant sources.

Implication for weight. Strong for these lessons:

• controls on new use leave the installed stock as the long-term source (pp. 66, 69–72);
• persistent agents shift harm to distant, future and non-human parties (pp. 70, 72).

The recovery evidence adds a corrective the chapter left out. Restriction does work, but slowly and incompletely (Fig. 6.1).

Claim 8. Releases are “inevitable”, and incidents “will continue to occur” (p. 71)#

Original claim (p. 71).

• Despite regulation, releases “are inevitable”.
• The 1999 Belgian feed contamination was found only because levels caused chick oedema.
• Lower-level food contamination incidents “will continue to occur in future”.
• The chapter also criticises the Belgian toxicologists who judged population-level health effects unlikely, likening them to Sanford Brown in 1937.

Developments since 2001.

What happened in Belgium in 1999.

• About 100 litres of PCB oil entered recycled fat used to make feed. That is about 50 kg as the seven marker congeners, or roughly 150 kg of total PCBs. The congener pattern matched Aroclor 1260/1254 (Monsanto’s trade name for PCB mixtures).
• Poultry poisoning was noticed in February 1999, but the source and extent were identified only in May (Bernard et al., Environ Res 88:1–18, January 2002). The chapter’s point that detection depended on acute effects in animals is confirmed.
• The consequences included Belgian PCB norms, EU-harmonised dioxin limits for feed and food in 2002, routine national monitoring, and a new federal food safety agency (Covaci et al., Environ Toxicol Pharmacol 25:164–170, March 2008; abstract read).

Incidents recurred, including through the same transformer-oil pathway. EFSA (2018) and the Commission (COM(2010)562) record:

• Germany, 2003: bread crumbs for feed dried using painted wood.
• Ireland, 2008: bakery waste for pig feed dried with burner fuel “apparently contaminated with PCB transformer oil”. This incident was found through routine monitoring of PCBs in pig meat, not through signs of disease.
• Germany, 2011: beet pulp dried with coal containing plastic.
• Free-range laying hens: incidents traced to recycled building materials.

How the rules changed.

• Maximum levels for non-dioxin-like PCBs in food (Regulation (EU) 1259/2011, 2 December 2011).
• Mandatory dioxin and dioxin-like PCB testing, and approval of establishments, for fats and oils used in feed (Regulation (EU) 225/2012, 15 March 2012).
• Commission recommendations on action levels in 2011, 2013 and 2014.

The Belgian risk assessment the chapter criticised.

• Later evidence supports the Belgian toxicologists at the level of the general population.
• Plasma from 232 Belgian blood donors taken before and after the incident showed an unchanged total TEQ: 22.9 against 23.1 pg WHO-TEQ/g fat. Two incident-specific furan congeners rose slightly (Debacker et al., Chemosphere 67:S217–S223, April 2007).
• Bernard and Fierens concluded that the incident was too limited in time and scale to raise the population’s body burden (Int J Toxicol 21:333–340, September 2002).
• A competing estimate, whose co-authors include the source of the chapter’s Belgian contamination data (Hens), put extra cancers at between 40 and 8,000. That is a range of more than two orders of magnitude (van Larebeke et al., EHP 109:265–273, March 2001).
• Neither set of data tested the chapter’s specific concern about fetal effects at background levels.

Verdict: held up on the inevitability and recurrence of releases. The Belgian-critique element is weakened, and the concern about detection has been partly addressed by institutionalised monitoring.

Implication for weight.

• Strong for “legacy stocks re-enter the food chain through reuse and waste pathways” (pp. 66, 71). The Irish case repeated the transformer-oil route almost exactly.
• Moderate for “detection favours acute, visible events” (p. 71). That was true in 1999 and less so once surveillance became routine, which shows how monitoring changes after a crisis.
• The chapter’s rhetorical comparison to 1937 should not be carried forward.

Claim 9. Action taken only at high proof; the 1970s “half measure” (p. 72)#

Original claim (p. 72).

• “At almost every stage government action was taken only when there was a high level of scientific proof.”
• The 1970s restrictions left existing uses and clean-up untouched, “probably” because of cost, even though alternatives were available.

Developments and evidence.

Evidence for the claim.

• EPA, 1979. EPA’s final rule allowed existing equipment holding about 578 of 750 million pounds of PCBs to stay in service because immediate replacement “would be prohibitively expensive” (EPA, 19 April 1979). This directly confirms the chapter’s suggested reason, cost.
• The EC, 1985. The Community’s 1985 amendment said a more general ban was “not feasible at this stage”. It allowed PCBs in existing equipment until the end of its service life (Directive 85/467/EEC, 1 October 1985).
• The courts, 1980. The D.C. Circuit found “no substantial evidence” for EPA’s classification of transformers and capacitors as “totally enclosed”, or for its 50 ppm regulatory cut-off, and sent both back to the agency. It noted that EPA had itself found that any exposure could cause adverse effects. It also held that Congress had not designed section 6(e) of the Toxic Substances Control Act (TSCA) to regulate PCBs already in the environment (EDF v. EPA, 636 F.2d 1267, 30 October 1980; text via Caselaw Access Project). Here the courts pushed the regulator beyond where it had chosen to stop.
• Deadlines kept slipping for the installed stock: the EU’s 2010 deadline, then 2025 in the EU and globally (Claim 7). US rules still permit PCB transformers.

Evidence that complicates it.

• Several early actions rested on evidence of persistence and bioaccumulation rather than proof of human harm at environmental levels:
• Monsanto’s withdrawal from open uses in 1970–72;
• the OECD decision of 1973 (p. 66);
• the EC’s 1976 Directive, which restricted PCBs with the stated aim of “moving gradually towards a complete ban” (Directive 76/769/EEC, 27 July 1976; annex conditions not examined here);
• the Toxic Substances Control Act of 1976, a statutory ban on manufacture.

By the evidentiary standards of the time, these were fairly precautionary. - Keeping closed uses reflected an explicit risk–risk judgement about fire and explosion. This appears in the 1969 plan and in the 1972 US Interdepartmental Task Force as quoted by Monsanto. The chapter does not engage with it. - The chapter also treats “production” loosely. In the US, voluntary withdrawal and statutory bans ended new production in 1977 (see Claim 10). Global production continued elsewhere into the 1980s, as the chapter itself notes (p. 67).

Verdict: partly held up. The “half measure” and the cost explanation are well supported by primary regulatory records. The claim that action waited for “high proof” at almost every stage is too sweeping. Early restrictions on open uses and new production were taken at moderate proof. The lag concentrated on the installed stock, where cost and fire-safety trade-offs dominated.

Implication for weight. Moderate. The better-supported lesson is that action was sequenced by cost:

1. cheap open uses first;
2. new production next;
3. the expensive installed stock last.

Deadlines for the stock then slipped repeatedly over four decades (pp. 66, 72). A second lesson is that courts and legislatures sometimes pushed agencies beyond their chosen stopping points.

Claim 10. “Some 100 years”; production end dates; Jensen’s 1969 identification (pp. 67, 69, 73)#

Original claims.

• “Some 100 years after the first serious adverse effects had been documented, closure had finally been reached” around 1995 (p. 69). The count starts with 1899 chloracne.
• Production ended “in the United Kingdom in 1978 and in the United States in 1979” (p. 67).
• Table 6.1 (p. 73): Jensen detected unknown molecules in 1966, but “only in 1969 was he able to demonstrate that they were PCBs”. The text (p. 64) says it took “two further years”.

Developments and evidence.

The 1899 starting point.

• Chloracne was “first described by Herxheimer in 1899” (Passarini et al., Dermatology 221:63–70, 2010). The cases involved chlorinated compounds in general, three decades before commercial PCBs (1929).
• A 1943 review traces the attribution of chloracne first to chlorinated naphthalenes (1918) and then to chlorinated naphthalenes and diphenyls in the 1930s (Greenburg 1943).
• Effects specific to PCB workers were observed at Swann before 1934 and published in 1936 (Claim 3).
• The interval from PCB-specific warnings to the global agreement of 1995 is therefore about 60 years, not 100.

“Closure” in 1995 was not final.

• IARC reclassified PCBs, including dioxin-like PCBs, as Group 1 human carcinogens in February 2013 (Lauby-Secretan et al., Lancet Oncol 14:287–288, April 2013).
• EFSA cut the TWI sevenfold in 2018, and JECFA evaluated non-dioxin-like PCBs for the first time in 2016 (Claim 4).
• The seriousness of the threat has been reaffirmed and extended. Some specific effects, such as neurodevelopment at background levels, have become less certain.

US production ended in 1977.

• EPA’s April 1979 release says PCBs were manufactured “from 1929–1977” and are “no longer being produced in this country” (EPA 1979).
• Monsanto says it ceased production and distribution in 1977, with its last inventory to be shipped by 31 October 1977 (Monsanto 1979). The Washington Supreme Court agrees (Erickson, 2025).
• The 1979 date is when the TSCA manufacturing ban took effect.
• EPA’s site profile says the Anniston plant itself made PCBs from 1929 until 1971 (EPA Anniston PCB site profile, archived 27 December 2016).

UK production ended in 1977. Monsanto’s Newport plant made PCBs between 1960 and 1977, according to trade press. This is a secondary source, not checked against a primary UK record (ENDS Report, c. 2007, archived).

Jensen identified PCBs by December 1966.

• New Scientist reported his identification of PCBs in pike, an eagle, and his own family’s hair on 15 December 1966 (“Report of a New Chemical Hazard”; copy in Monsanto’s files: Toxic Docs). A Monsanto memo of 31 July 1968 circulated it internally (Toxic Docs).
• Monsanto’s own records date the Swedish identification to late 1966:
• the 1969 plan says Widmark and Jensen “in November 1966, announced and confirmed” PCB in fish, birds and eggs (plan copy, Toxic Docs);
• the 1979 account says “late 1966”.
• The 1969 Nature paper was the full publication of the data.
• Francis quotes Jensen as having first seen the unknown peaks in 1964 (secondary source).
• Jensen’s own 2021 retrospective does not give dates (Ambio 50:525–526, March 2021).

Production totals. Estimates of total global production:

The chapter’s 1.5 million t (excluding the USSR and China) sits at the top of this range.

Verdict: weakened. The “100 years” framing is inflated by about 40 years if PCB-specific evidence is the test. All three dates are wrong: US and UK production both ended in 1977, and PCBs were identified in 1966.

Implication for weight. The long-latency story still holds in outline:

• about 30 years from the first PCB-specific worker warnings to discovery in the environment;
• about 60 years to global agreement;
• decades more to disposal.

Exact intervals from this chapter should not be reused without these corrections.

What this means for the section’s lessons (technology-neutral)#

1. Controls on new use leave the installed stock as the long-term source of harm (pp. 66, 69–72). Strengthened. - The EU’s 2010 deadline slipped. - The Stockholm 2025 target was off track on the Secretariat’s own assessments, and 2028 targets remain. - One major producer country still permits lifetime use. - The largest remaining EU emission source is the old equipment.
2. The product tested is not the exposure people and ecosystems receive (p. 67). Held up, and built into regulatory practice (EPA’s tiered slope factors).
3. Private acknowledgement can coexist with public calls for “more research”, and liability discourages visible admission (p. 65). Held up on primary evidence. - Reuse should cite the 1969 plan and the March 1969 letter directly. - It should also note that the firm’s reasons were plural, commercial and legal, and framed around fire safety.
4. Warnings confined to a closed circle (pp. 64, 71). Needs reframing for the 1930s. - The evidence was published, and public-health officials were present. - What limited its reach was framing: an occupational hazard, controllable by ventilation, with no concept of persistence in the environment. - Private knowledge diverging from public statements is best evidenced for the late 1960s.
5. Action sequenced by cost (pp. 66, 72). Held up. EPA in 1979 (“prohibitively expensive”) and the EC in 1985 (“not feasible”) say so in their own words.
6. Risk–risk trade-offs. An omission. Fire and explosion risks were central to the 1969–1972 decisions to keep closed uses. Lessons drawn from this chapter should add the trade-off and show how it was later resolved by substitution within a few years.
7. Heterogeneous classes and mechanism (pp. 66–70). Mixed. The lesson that congeners differ and that lumping them breeds confusion holds. The chapter’s own attribution to non-dioxin-like congeners shows the opposite risk: over-reading mechanism from correlated exposures.
8. Detection favours acute, visible events (p. 71). True for 1999, partly superseded. Monitoring was institutionalised after the crisis and caught the 2008 Irish incident at a sub-clinical stage.
9. Recovery after restriction (Fig. 6.1). Strengthened but incomplete. Baltic seals and eagles recovered. Top marine predators remain at risk, and environmental declines are levelling off.
10. Advocacy exceeded evidence (p. 72). Confirmed. The chapter’s paediatric claims went beyond the evidence then and have not been borne out since. The 2013 update extends the same pattern to diabetes and epigenetics. This fits the digest’s caution about the authors’ standpoint.

Corrections to carry forward#

• p. 65. “profits to cease and liability to soar” is not in the 1969 plan. It is Francis’s paraphrase. The primary text reads “We would be admitting guilt by our actions”, and the plan gives several other reasons as well.
• p. 65 and references. “Francis 1998” is a web reprint of Eric Francis, “Conspiracy of Silence”, Sierra, September/October 1994.
• p. 64. Of the three Halowax deaths, one involved a mixture containing chlorinated diphenyl, at about 10%. The “mob hysteria” remark appears in the published 1937 journal record.
• p. 64 and Table 6.1 (p. 73). Jensen’s identification of PCBs was public by December 1966, not 1969.
• p. 67. US production ended in 1977; 1979 is when the TSCA ban took effect. UK production also appears to have ended in 1977 (trade-press source).
• p. 69. “Some 100 years” counts from a class-level 1899 warning. For PCB-specific evidence, about 60 years to 1995.

Sources#

Primary historical documents (Toxic Docs litigation archive) - Monsanto, “PCB Environmental Pollution Abatement Plan” (November 1969; Bates WATER PCB-00049709–29, NEV 021436–56): https://www.toxicdocs.org/d/kZv8Ew16Rqo0YxLNrQwg8y9J - Monsanto, “PCB Environmental Pollution Abatement Plan”, rough draft dated 10 November 1969 (MONS 035330 series): https://www.toxicdocs.org/d/5LOjyE97dEnzo8yyxLeMQw2Vz - Monsanto, “PCB Environmental Pollution Abatement Plan”, further copy (legible “November 1966” passage on Widmark and Jensen): https://www.toxicdocs.org/d/Edkm24ZvdJeLEXQdQ1xYoQ5j0 - Monsanto customer letter, E. P. Wheeler, 3 March 1969: https://www.toxicdocs.org/d/ymnxjL9YKpQ9Z6NYevoEK5Mpn - Drinker CK, Warren MF, Bennett GA, “The problem of possible systemic effects from certain chlorinated hydrocarbons”, J Ind Hyg Toxicol 19(7):283–311, September 1937, with the symposium discussion: https://www.toxicdocs.org/d/mq47bDRVyO6OKz05VnEk3n2X0 - Key to participants at the 1937 meeting (plaintiff-side annotation of the published record): https://www.toxicdocs.org/d/Yrg6go5r8NzY93MDE6Jz3qQzN - “Report of a New Chemical Hazard”, New Scientist, 15 December 1966 (copy in Monsanto files): https://www.toxicdocs.org/d/evbbezoy4eG2M3njB3DMRN0ry - Monsanto internal memo circulating the New Scientist item, 31 July 1968: https://www.toxicdocs.org/d/37QggOxjjKb5EJnxRzjLmgrNJ - Monsanto, report on the history of chlorinated biphenyl production, sales and health information, 16 November 1979: https://www.toxicdocs.org/d/ExvB3zp2y34DqqN1q1eyddB5N - Greenburg L, “Safety measures for use of chlorinated naphthalenes and diphenyls in industry”, Industrial Bulletin (NY State) 22(10), October 1943: https://www.toxicdocs.org/d/KRRNGg5egamG1VM314pekye2w

Secondary sources used by the chapter, and scholarship - Francis E, “Conspiracy of Silence” (Sierra, September/October 1994; web reprint cited by EEA as Francis 1998), Wayback capture of 5 December 2000: http://web.archive.org/web/20001205190900/http://www.planetwaves.net:80/silence2.html - Markowitz G, Rosner D, “Monsanto, PCBs, and the creation of a ‘world-wide ecological problem’”, J Public Health Policy 39(4):463–540, November 2018 (abstract only read): https://doi.org/10.1057/s41271-018-0146-8 - Jones JW, Alden HS, “An acneform dermatergosis”, Arch Dermatol Syphilol 33(6):1022, June 1936 (bibliographic record only): https://doi.org/10.1001/archderm.1936.01470120073010 - Jensen S, “Afterthoughts from an environmental pollution discovery”, Ambio 50:525–526, March 2021: https://doi.org/10.1007/s13280-020-01453-2 - Breivik K et al., “Towards a global historical emission inventory for selected PCB congeners… 1. Global production and consumption”, Sci Total Environ 290:181–198, 2002: https://doi.org/10.1016/S0048-9697(01)01075-0 - Passarini B et al., “Chloracne: still cause for concern”, Dermatology 221:63–70, 2010: https://doi.org/10.1159/000290694

EEA - EEA, Late lessons from early warnings: science, precaution, innovation (Report 1/2013), Annex 3, PCB update by J. G. Koppe, p. 727 (archived annex PDF): http://web.archive.org/web/20250911210532/https://www.eea.europa.eu/publications/late-lessons-2/late-lessons-2-full-report/late-lessons-2-annexes/at_download/file (landing page: https://www.eea.europa.eu/en/analysis/publications/late-lessons-2)

Risk assessments (WHO, EFSA, IARC, US EPA) - EFSA CONTAM Panel, dioxins and dioxin-like PCBs in feed and food, EFSA Journal 16(11):5333, November 2018: https://doi.org/10.2903/j.efsa.2018.5333 - EFSA CONTAM Panel, update of the risk assessment of dioxins and dioxin-like PCBs in feed and food, EFSA Journal, 10 June 2026 (full opinion not read): https://doi.org/10.2903/j.efsa.2026.10103 - EFSA, plain-language summary of the 2026 update, 10 June 2026: https://www.efsa.europa.eu/en/plain-language-summary/update-risk-assessment-dioxins-and-dioxin-pcbs-feed-and-food - EFSA news, “EFSA confirms health concern from dietary exposure to dioxins and dioxin-like PCBs in Europe”, 10 June 2026: https://www.efsa.europa.eu/en/news/efsa-confirms-health-concern-dietary-exposure-dioxins-and-dioxin-pcbs-europe - DeVito M et al., “The 2022 World Health Organization reevaluation of human and mammalian toxic equivalency factors for polychlorinated dioxins, dibenzofurans and biphenyls”, Regul Toxicol Pharmacol 146:105525, January 2024: https://www.sciencedirect.com/science/article/pii/S0273230023001939 - JECFA, Safety evaluation of certain food additives and contaminants, Supplement 1: Non-dioxin-like PCBs, WHO Food Additives Series 71-S1, 15 July 2016: https://iris.who.int/handle/10665/246225 - van den Berg M et al., “WHO/UNEP global surveys of PCDDs, PCDFs, PCBs and DDTs in human milk and benefit–risk evaluation of breastfeeding”, Arch Toxicol 91:83–96, January 2017: https://doi.org/10.1007/s00204-016-1802-z - Lauby-Secretan B et al. (IARC), “Carcinogenicity of polychlorinated biphenyls and polybrominated biphenyls”, Lancet Oncol 14:287–288, April 2013: https://doi.org/10.1016/S1470-2045(13)70104-9 - US EPA IRIS, Polychlorinated biphenyls, chemical assessment summary (carcinogenicity last revised 1 October 1996): https://iris.epa.gov/static/pdfs/0294_summary.pdf

US regulation and courts - US EPA press release, “EPA Bans PCB Manufacture; Phases Out Uses”, 19 April 1979: https://www.epa.gov/archive/epa/aboutepa/epa-bans-pcb-manufacture-phases-out-uses.html - 40 CFR 761.30, Authorizations (eCFR, as of 1 September 2026): https://www.ecfr.gov/current/title-40/chapter-I/subchapter-R/part-761/subpart-B/section-761.30 - Environmental Defense Fund v. EPA, 636 F.2d 1267 (D.C. Cir., 30 October 1980): https://www.courtlistener.com/opinion/385701/environmental-defense-fund-inc-v-environmental-protection-agency/ (text: https://static.case.law/f2d/636/cases/1267-01.json) - Erickson v. Pharmacia LLC, Washington Supreme Court No. 103135-1, 30 October 2025 (amended 21 January 2026): https://www.courts.wa.gov/opinions/pdf/1031351.pdf - US EPA, Anniston PCB Site (Monsanto Co), Superfund site profile (archived 27 December 2016): https://web.archive.org/web/20161227015252/https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0400123 - Washington Attorney General, “Monsanto to pay record $95 million to end Ferguson’s lawsuit over PCBs”, 24 June 2020: https://www.atg.wa.gov/news/news-releases/monsanto-pay-record-95-million-end-ferguson-s-lawsuit-over-pcbs - OPB, “Oregon reaches nearly $700M settlement with Monsanto over PCB contamination”, 15 December 2022 (secondary): https://www.opb.org/article/2022/12/15/oregon-settlement-monsanto-pcb-contamination-attorney-general-ellen-rosenblum/ - Michigan Attorney General, “AG Nessel Secures $108 Million Settlement with Monsanto for Impacts from Toxic PCBs”, 18 May 2026 (access error; not read directly): https://www.michigan.gov/ag/news/press-releases/2026/05/18/ag-nessel-secures-settlement-with-monsanto-for-impacts-from-toxic-pcbs - Great Lakes Now, “Michigan gets $108M in Monsanto settlement to clean up PCB contamination”, 20 May 2026 (secondary): https://www.greatlakesnow.org/2026/05/20/michigan-gets-108m-in-monsanto-settlement-to-clean-up-pcb-contamination/ - Bayer, “Managing and Mitigating the U.S. PCB Litigation Risk” (company page, undated, consulted September 2026; access error, wording from search-result extracts): https://www.bayer.com/en/resolving-us-pcb-litigation - Law360, jury verdict for Monsanto in Los Angeles PCB–lymphoma trial, 22 May 2014 (secondary; archived): https://web.archive.org/web/20150612002836/http://www.law360.com/articles/540961/jury-finds-monsanto-pcbs-not-to-blame-for-cancer-cases - St Louis Post-Dispatch, “Monsanto prevails in PCB lawsuit”, 7 July 2015 (secondary; archived): https://web.archive.org/web/20161205221536/http://www.stltoday.com/business/local/monsanto-prevails-in-pcb-lawsuit/article_9a5699ef-f545-5ad2-8f67-624f78fb10b7.html

EU law, reports and judgments - Council Directive 76/769/EEC, 27 July 1976: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:31976L0769 - Council Directive 85/467/EEC, 1 October 1985: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:31985L0467 - Council Directive 96/59/EC on the disposal of PCBs/PCTs, 16 September 1996: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:31996L0059 - CJEU: C-46/01 Commission v Italy, 27 February 2002 (https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62001CJ0046); C-177/01 Commission v France, 6 June 2002 (https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62001CJ0177); C-47/01 Commission v Spain, 3 October 2002 (https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62001CJ0047); C-174/01 Commission v Luxembourg, 5 December 2002 (https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62001CJ0174); C-83/02 Commission v Greece, 5 June 2003 (https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62002CJ0083); C-508/06 Commission v Malta, 29 November 2007 (https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62006CJ0508) - Commission, third progress report on the Community Strategy for dioxins, furans and PCBs, COM(2010)562, 15 October 2010: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52010DC0562 - Commission Regulation (EU) 1259/2011 (maximum levels for dioxins, dioxin-like and non-dioxin-like PCBs in food), 2 December 2011: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32011R1259 - Commission Regulation (EU) 225/2012 (approval and dioxin testing for feed fats and oils), 15 March 2012: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32012R0225 - Regulation (EU) 2019/1021 on persistent organic pollutants, 20 June 2019, Annex I: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32019R1021 - Commission Staff Working Document, EU Implementation Plan for the Stockholm Convention, SWD(2021)201, 22 July 2021: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52021SC0201 - Commission Staff Working Document, Union Synthesis Report on the application of Regulation (EC) No 850/2004, SWD(2022)291, 16 September 2022: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52022SC0291

Stockholm Convention - PCB overview page (Secretariat): https://chm.pops.int/Implementation/IndustrialPOPs/PCBs/Overview/tabid/273/Default.aspx - Progress towards PCB elimination page (Secretariat): https://www.pops.int/Implementation/IndustrialPOPs/PCB/ProgresstowardsPCBelimination/tabid/8304/Default.aspx - UNEP/POPS/COP.11/INF/11, “Report on progress towards the elimination of polychlorinated biphenyls”, 18 April 2023: https://chm.pops.int/Portals/0/download.aspx?d=UNEP-POPS-COP.11-INF-11.English.pdf - UNEP/POPS/COP.12/INF/11, “Revised strategy for Parties to meet the 2025 and 2028 goals on PCBs”, 3 February 2025: https://chm.pops.int/Portals/0/download.aspx?d=UNEP-POPS-COP.12-INF-11.English.pdf - Decision SC-12/3, Polychlorinated biphenyls (COP-12, Geneva, 28 April–9 May 2025): https://chm.pops.int/Portals/0/download.aspx?d=UNEP-POPS-COP.12-SC-12-3.English.pdf - UN Treaty Collection, status of the Stockholm Convention: https://treaties.un.org/Pages/ViewDetails.aspx?src=TREATY&mtdsg_no=XXVII-15&chapter=27&clang=_en - GEF, “Accelerating progress with electricity upgrades” (Global Elimination Program for PCBs), 24 April 2025: https://www.thegef.org/newsroom/feature-stories/accelerating-progress-electricity-upgrades - African Development Bank, “Abidjan Workshop Advances Regional Drive to Meet Stockholm Convention PCB Elimination Targets”, early 2026 (access error; details from search-result extracts): https://www.afdb.org/en/news-and-events/abidjan-workshop-advances-regional-drive-meet-stockholm-convention-pcb-elimination-targets-90567

Epidemiology - Schantz SL et al., EHP 111:357–576, March 2003: https://doi.org/10.1289/ehp.5461 - Jacobson JL, Jacobson SW, J Pediatr 143:780–788, December 2003: https://doi.org/10.1067/S0022-3476(03)00577-8 - Vreugdenhil HJ et al., Dev Med Child Neurol 46:398–405, June 2004: https://doi.org/10.1017/S0012162204000647 - Weisglas-Kuperus N et al., Toxicol Lett 149:281–285, April 2004: https://doi.org/10.1016/j.toxlet.2003.12.039 - Gray KA et al., Am J Epidemiol 162:17–26, July 2005: https://doi.org/10.1093/aje/kwi158 - Heilmann C et al., PLoS Med 3:e311, August 2006: https://doi.org/10.1371/journal.pmed.0030311 - Dallaire F et al., EHP 114:1301–1305, August 2006: https://doi.org/10.1289/ehp.8683 - Boucher O et al., EHP 117:7–16, January 2009: https://doi.org/10.1289/ehp.11294 - Gascon M et al., Environ Int 52:51–65, February 2013: https://doi.org/10.1016/j.envint.2012.11.005 - Fång J et al., Environ Int 60:224–231, October 2013: https://doi.org/10.1016/j.envint.2013.08.019 - Gascon M et al., Epidemiology 25:544–553, July 2014: https://doi.org/10.1097/EDE.0000000000000097 - Forns J et al., Int J Epidemiol 47:1082–1097, August 2018: https://doi.org/10.1093/ije/dyy052 - Berghuis SA et al., Environ Int 121:13–22, December 2018: https://doi.org/10.1016/j.envint.2018.08.030 - Xu K et al., Environ Pollut 336:122439, November 2023: https://doi.org/10.1016/j.envpol.2023.122439 - Balalian AA et al., Environ Res 252:118912, July 2024: https://doi.org/10.1016/j.envres.2024.118912

Wildlife and environment - Heaton SN et al., Arch Environ Contam Toxicol 28(3), April 1995: https://doi.org/10.1007/BF00213111 - Bursian SJ et al., Environ Toxicol Chem 25:1533–1540, June 2006: https://doi.org/10.1897/05-406R.1 - Fuchsman PC et al., Arch Environ Contam Toxicol 54:130–144, January 2008: https://doi.org/10.1007/s00244-007-9001-9 - Bergman A, APMIS 107:270–282, 1999: https://doi.org/10.1111/j.1699-0463.1999.tb01554.x - Helander B et al., Ambio 37:425–431, September 2008: https://doi.org/10.1579/0044-7447(2008)37[425:URAESM]2.0.CO;2 - Jepson PD et al., Sci Rep 6:18573, 14 January 2016: https://doi.org/10.1038/srep18573 - Desforges JP et al., Science 361:1373–1376, 28 September 2018: https://doi.org/10.1126/science.aat1953 - ENDS Report, on Brofiscin Quarry and Monsanto’s Newport PCB production (trade press, c. 2007; archived): https://web.archive.org/web/20190804122813/https://www.endsreport.com/article/1571135/agency-legal-battle-clean-pcb-waste-dump

The 1999 Belgian incident - van Larebeke N et al., EHP 109:265–273, March 2001: https://doi.org/10.1289/ehp.01109265 - Bernard A et al., Environ Res 88:1–18, January 2002: https://doi.org/10.1006/enrs.2001.4274 - Bernard A, Fierens S, Int J Toxicol 21:333–340, September 2002: https://doi.org/10.1080/10915810290096540 - Debacker N et al., Chemosphere 67:S217–S223, April 2007: https://doi.org/10.1016/j.chemosphere.2006.05.101 - Covaci A et al., Environ Toxicol Pharmacol 25:164–170, March 2008: https://doi.org/10.1016/j.etap.2007.10.003