LL2-03 — Ch3 Lead in petrol ‘makes the mind give way’#

Late lessons from early warnings: science, precaution, innovation (EEA Report No 1/2013), Part A “Lessons from health hazards”. Report pages 46–75 (PDF pages 48–77). Body text pp. 46–71; references pp. 72–75.

Read in full from the text extract (every page marker, PDF 48–77). Visually checked in the PDF: chapter opener (p. 46), Box 3.1 (p. 47), Box 3.5 (p. 51), Panel 3.1 (p. 54), Box 3.10 and Box 3.11 (p. 62), Figure 3.1 (p. 64), Figure 3.2 (p. 66), p. 69 and Table 3.1 (p. 71). The extraction is faithful for body text. Figures are summarised from the rendered pages. Author affiliations come from the report’s author annex (pp. 689–699), because the chapter itself states none.

Authors and standpoint#

Chapter authors: Herbert Needleman and David Gee (p. 46).

• Herbert Needleman is described in the annex as a paediatrician and child psychiatrist at the University of Pittsburgh Medical Center (p. 695). The bio adds that his “research and advocacy were instrumental in federal regulations banning lead from gasoline and paint” (p. 695). He is a protagonist in the story he co-writes. The chapter describes his own 1979 study as “seminal” (p. 60) and reports industry’s “character assassination” of him (p. 61). Box 3.9 is a first-person anecdote of his (p. 61), and he appears among the vindicated “early warning” scientists in the lessons (pp. 70–71). His earlier historical essays (Needleman 1997, 1998, 2000) are among the chapter’s main sources.
• David Gee trained in economics and politics. He worked for trade unions and NGOs from 1974, is a former Director of Friends of the Earth UK, and has worked at the EEA since 1995. He originated the Late lessons project and edited both volumes (p. 689; front matter PDF p. 7).
• Acknowledgement (p. 46, fn 1): the authors thank Gerald Markowitz and David Rosner for “their detailed history” in Deceit and denial (2002). Much of the 1920s–1960s narrative follows Markowitz and Rosner and Needleman’s essays. The archival letters (National Archives Record Groups 70 and 90) seem to be cited through these secondary histories. The chapter does not say whether the authors consulted them directly.
• Evident stance: strongly pro-precaution and critical of industry. The chapter is sympathetic to “early warning” scientists and frames the case as “‘Progress’ or precaution?” (section title, p. 52). It is written as a morality tale with clear heroes (Henderson, Hamilton, Patterson, Needleman) and antagonists (Midgley, Kettering, Howard, Kehoe, Ethyl/GM/DuPont/Standard Oil, the captured public health establishment). The tone is often polemical, for example “Apparently, poisoning a technology was more important than poisoning people” (p. 60) and “plumbing new depths” (p. 57).
• The front matter says case study chapters were “peer reviewed from a broad range of perspectives” (PDF p. 7). No industry or regulator reply appears in this chapter.

Panels (all four broadly support the chapter; none dissents, and none speaks for industry or regulators):

Section-by-section notes#

Chapter summary box (p. 46)#

• Scope: lead in petrol, 1925–2005. Leaded petrol was first widely marketed in the US, spread worldwide, then phased out from the 1970s. In Europe, “the Aarhus Protocol (UNECE, 1998) initiated the phase-out of leaded petrol in the period 1998–2005” (p. 46). (The reference list entry for UNECE 1998, p. 75, is the Aarhus Convention on access to information, which looks like a citation error. See the bias check below.)
• Core claims:
• Lead’s neurotoxicity was known since Roman times.
• At the 1925 “one day trial” many experts warned of health impacts.
• An “equally effective alcohol additive which was assessed by experts to be cleaner” was available, yet “the leaded route to fuel efficiency was chosen in the US and then exported to the rest of the world” (p. 46).
• For decades “virtually no independent research was carried out and the main source of information was industry and industry-sponsored researchers”. Only in the 1960s–70s did independent scientists show that body burdens were not “normal” but “hundreds of times higher than before the industrial revolution” (p. 46).
• At the mid-1970s peak, leaded petrol released “about 200 000 tonnes of lead into the atmosphere annually in both the US and Europe”. After the phase-out, children’s blood lead “quickly fell, in line with the decrease in air concentrations” (p. 46).
• Continuing relevance: nearly all countries had phased out by 2012, but soil and sediment lead remains high, and e-waste causes elevated blood lead (p. 46).

Epigraph and 3.1 Introduction (p. 47)#

• Opens with Yandell Henderson’s 1925 prediction that “conditions will grow worse so gradually, and the development of lead poisoning will come on so insidiously … that leaded petrol will be in nearly universal use … before the public and the government awakens to the situation” (USPHS 1925; p. 47).
• “There were 50–70 years between Henderson’s prescient early warning … and its phase out in the US and then Europe in the 1970s and mid-1980s” (p. 47). The history is “rich in lessons about the science, economics, and politics of identifying and controlling the hazards of toxic substances” (p. 47).
• The focus is children (Box 3.1). The longer histories of lead in pots and paint “weave in and out of the leaded petrol saga, both complicating the search for the causes of lead poisoning in children” and creating a persistent environmental stock (p. 47). This is an early acknowledgement of multi-source attribution difficulty.
• Adults are affected too: hypertension, heart disease and kidney disease (Navas-Acien et al. 2007; EFSA 2010) (p. 47).
• The lessons are claimed to be “relevant for controlling most toxic chemicals” and for the millions still exposed (pp. 47–48).

Box 3.1 Children and lead: health impacts (source WHO 2010; p. 47) - High chronic exposure causes coma, convulsions and death. Survivors of acute poisoning are typically left with “grossly obvious mental retardation and behavioural disruption”. - At lower exposures “previously considered safe” the effects include diminished cognition, shortened attention span, disruptive behaviour, dyslexia, ADD, hypertension, renal impairment, immunotoxicity and reproductive toxicity. They are “for the most part … permanent and largely untreatable”. - Ten exposure sources are listed: mining, paint, can solder, ceramic glazes, water pipes and solder, products (traditional medicines, cosmetics, toys), incineration, e-waste, the food chain via soil, and legacy sites. - Acute poisoning persists in low-income countries and marginalised populations. - Neurobehavioural damage occurs at “5 μg/dl and even lower”. “There appears to be no threshold level below which lead causes no injury to the developing human brain” (p. 47). - “The biology of childhood lead poisoning is similar everywhere”, so studies transfer across countries (p. 47).

Box 3.2 Ancient lead poisoning (p. 48)#

• The Greenland ice core (Hong, Candelone et al. 1994, with Patterson as co-author) shows lead at four times natural values between c. 500 BC and 300 AD.
• Cumulative fallout during those eight centuries was “as high as 15 % of that caused by use of lead alkyl additives in petrol since the 1930s” (p. 48).

3.2 Lead toxicity: some early warnings (pp. 48–49)#

• Dioscorides (first century AD, physician in Nero’s army) wrote in Materia Medica that “Lead makes the mind give way” (p. 48). This is the source of the chapter title. Exposure came from glazed pots and from lead in winemaking (“sugar of lead” sweetened wine). Toddlers found that leaded paint “tasted like lemon drops” (p. 48).
• 1723: a US law (colony not named) banned leaden “worms” in rum distilling to prevent the “dry gripes”. The £100 penalty was split between the poor and the informer, “an early attempt to reward and protect ‘whistleblowers’” (cross-ref Chapter 24) (p. 48).
• 18th-century Devon colic: Sir George Baker traced it to lead in cider presses (Baker 1768). He “was condemned by the clergy, by mill owners and by fellow physicians: cider was Devon’s main export” (p. 48). Baker “suffered the fate of many ‘early warning’ scientists whose inconvenient truths are not welcomed by supporters of the status quo” (p. 48).
• Footnote 2 (p. 48) cites Ibsen’s An enemy of the people on “shooting the messenger”, and Dr Hosokawa at Chisso (Minamata), who was told to suppress his discovery (cross-ref Chapter 5).
• Josiah “Wedgewood” (sic) asked the British government to extend the 1833 Factories Act to the potteries so that he could have a “level playing field” with less scrupulous competitors. Opposition from other manufacturers meant a wait of “some 30 years” until the Potteries Regulations of 1867 (p. 48). This matters because it shows a responsible firm seeking regulation to neutralise the competitive advantage of laggards. (Chronology problem: Josiah Wedgwood the founder died in 1795. See the bias check.)
• Box 3.3 (p. 49): Benjamin Franklin on the “dangles” (wrist drop) and “dry gripes” in painters, potters and plumbers. He solved the puzzle of lead poisoning in stonecutters and soldiers by tracing their hidden exposures, “an early example of the value of detailed job and life exposure histories”. Franklin: “you will observe with concern how long a useful truth may be known and exist, before it is generally received and practiced on” (Franklin 1818, p. 49). (The box says “In 1818, Benjamin Franklin, while ambassador to France”. Franklin died in 1790, so 1818 is the publication date of his collected works.)
• Lead in pots, paints, pipes and toys grew in the 19th and 20th centuries. The poisoning was “insidious” and often appeared years after exposure (Oliver 1911, p. 49). The chapter notes that this “insidious” character would later be raised by some experts assessing TEL (p. 49).
• Box 3.4 Paint (p. 49):
• Children are more vulnerable because of in utero exposure, higher intake per unit weight, hand-to-mouth behaviour and nutritional deficiencies.
• 1892: an Australian doctor linked paint on verandas to poisoning in 10 children (Gibson 1904). 1914: the first US childhood death, from crib-rail paint.
• “By 1925 there was much scientific evidence … showing that infants and children were poisoned by lead in the paint”. Protection came indirectly, via painters’ occupational rules, with bans in Europe and Australia from 1909 to the 1930s.
• The US banned interior leaded paint only in 1971. Remediation and litigation costs in poorer US areas “still consume much time and money”.
• A UNEP/WHO 2010 survey of 10 countries found paint at 4 000 to nearly 40 000 ppm, against a US limit of 90 ppm. China reports many incidents (p. 49).

3.3 Lead in petrol 1922–1925: the early warnings of hazards to the public (pp. 50–52)#

3.3.1 Origins (p. 50) - “Until 1925, the principal source of toxic lead for the public was household paint” (p. 50). - GM, second to Ford in sales, faced engine “knock” in its new higher-performance Cadillac. Chief chemist Thomas Midgley (spelt “Midgely” in the main text), “who later invented the CFC chemicals” (cross-ref EEA 2001 Ch. 7), found an old German patent for tetraethyl lead (TEL) (p. 50). - “Alternatives to petroleum-based fuel, such as ethanol, were available. They were likely to be much less profitable, however, especially given the family and financial links between GM, the chemical company, DuPont, and Standard Oil” (p. 50). Pierre du Pont chaired the GM board and his brother Irénée ran DuPont. “In 1924 the three companies created the Ethyl Corporation of America to produce TEL” (p. 50). - Naming: “It was made clear at the outset that the word ‘lead’ was not to be used in the company name or sales literature: the little known term, ‘ethyl’ was used instead so as not to alarm the public” (p. 50). The authors call this “an early example of the censoring of sensitive words from the discussion of hazards”. Their parallel is asbestos, where “cancer” was replaced by “tumour” or “malignancy” at manufacturers’ request (p. 50).

3.3.2 Early warnings and “authoritative assertions” of safety (pp. 50–52) - TEL had been evaluated as a possible battlefield weapon in World War I. - 1922: Mansfield Clark, described by the chapter as “a professor of chemistry familiar with this work” (the wartime evaluation), warned the US Public Health Service (PHS) of a “serious menace to the public health”, because “on busy thoroughfares it is highly probable that the lead oxide dust will remain in the lower stratum” (p. 50). The reference list dates this as a memorandum of 11 October 1922 to Assistant Surgeon General Stimson, sent “through Acting Director, Hygenic Laboratory” (p. 73). - Surgeon General Hugh Cumming (spelt “Huge Cummings”) wrote to Pierre du Pont asking whether health effects had been considered, “since lead poisoning in human beings is of the cumulative type resulting frequently from the daily intake of minute quantities” (p. 50). - Midgley’s reply (Dec 1922): GM had given the question “very serious consideration … although no actual experimental data has been taken”, but was confident that “the average street will probably be so free from lead that it will be impossible to detect it or its absorption” (p. 50). - The authors’ generalisation: interested parties make “‘authoritative assertions’ about the absence of risk despite having little or no data”. “‘No evidence of harm’ is thereby mischaracterised as ‘evidence of no harm’. This approach to early warnings of potential harm is still common” (p. 50). - Footnote 3 (p. 50) extends this to mobile phones: there are “no studies in children” of the head cancer hazard, yet “it is widely asserted that there are no risks to children”. The EEA issued “early warnings” in 2007, 2009 and 2011 (cross-ref Ch. 21). This is advocacy on a live controversy embedded in a historical chapter. - Bureau of Mines study: GM paid the government’s Bureau of Mines to run animal experiments “within tight reporting constraints imposed by the Ethyl Corporation”. “Lead” was replaced by “ethyl” even in internal correspondence, and drafts went to Ethyl for “comments, criticism and approval” before publication (p. 50). The Bureau’s chief chemist objected, but his director said “it would not be so bad if the word lead were omitted as this term is apt to prejudice somewhat against its use” (p. 50). - Henderson (Yale physiology) had already turned down a GM invitation to study TEL and condemned “contractual gagging”. He declined the Bureau’s invitation too, calling GM funding “extremely unfortunate” given the “urgent need for an absolutely unbiased investigation”. He would investigate only “on the assumption that so terrible a poison as TEL should not be generally introduced until absolute proof was available that no danger to the public would be involved” (p. 51). His contracts with the Bureau of Mines and Standard Oil were then terminated: “an early example of the harassment of ‘early warning’ scientists” (p. 51). The chapter labels this harassment outright, but the only evidence it offers is timing (“Soon after”). No source on motive is given. - Worker deaths (Box 3.5, p. 51): - Ernest Oelgert at Standard Oil’s Bayway plant began hallucinating on 26 October 1924 and died days later. - Four more Bayway workers died over the next five days and 35 had severe neurological symptoms. At least six others died at DuPont Deepwater and GM Dayton. - New York Times journalists “uncovered more than 300 cases of lead poisoning at the Deepwater plant” despite “declared difficulties in getting the facts out of the companies and the hospitals”. Workers called the plant “the house of butterflies” (p. 51). - Media and bans: headlines such as “Mad gas claims third victim” and “Bar Ethyl gas as fifth victim dies” appeared. “New York State then banned the sale of leaded petrol” (p. 51). This put pressure on the PHS and industry to show that the workers’ risk did not extend to the public. Management blamed worker “carelessness” or said they “worked too hard” (p. 51). - Standard Oil had “confidently asserted” that no “perils existed in the use of this gas in automobiles” with no evidence. The day after the fifth death, the Bureau of Mines published animal results showing no public risk. On 1 November 1924 the NYT headline read “No peril to public after long experiments with motor exhausts” (p. 51). - The Bureau report was called “inadequate” by Cecil Drinker (Journal of Industrial Hygiene), David Edsall (Dean of Harvard Medical School) and others, including the Surgeon General. They cited too few animals and too short an exposure. The authors say “these are still features of some current toxicology that can result in underestimation of hazards” (cross-ref Ch. 26) (p. 51). - Hayhurst’s editorial: Emery Hayhurst (Ohio Department of Health) wrote an unsigned AJPH editorial saying observational evidence had “corroborated the statement of complete safety so far as the public health has been concerned”. “Few people knew” that he was a paid consultant to Ethyl and an adviser to the Bureau of Mines (Hayhurst 1924 letter identifying himself as “Consultant to Ethyl Gasoline Corporation”, p. 73) (p. 52). This is a documented undisclosed conflict of interest. - These “authoritative assurances” failed to quell concern. The Surgeon General “responded to requests for action from public health experts, who felt that both the public’s health and the probity of the PHS were at risk” (p. 52), and called a high-level conference: leaded petrol “is a public health question of extreme seriousness … if this product is actually causing slow poisoning” (p. 52). The conference was held in Washington on “24 May 1925” (p. 52).

3.4 “Progress” or precaution? (pp. 52–53, 56)#

3.4.1 The “one day trial” of the “gift of God” (pp. 52–53) - “Every major stakeholder was represented” (p. 52). - Industry’s four points (p. 52): 1. Leaded petrol was essential to America’s industrial progress. 2. All innovation entails risks. 3. Worker deaths were due to the men’s carelessness. 4. Street exposure differed from factory exposure, so the public was not at risk. - “No ‘innovation’ other than TEL was discussed at the meeting, despite the declared intention of the Surgeon General to spend two or three days discussing alternatives” (p. 52). The scare quotes signal that the authors contest the idea that TEL was the innovation. - The morning session went to industry: toxicologist Robert Kehoe first, “citing lack of evidence of risks to the public”, then eight other industry representatives (p. 52). - In the afternoon, public health representatives “try to shift the burden of proof back to their opponents” (p. 52): - Henderson: lead is “as serious a public health menace as infectious diseases”, and the insidious-spread prediction. “His claims proved to be prescient” (p. 52). - Edsall rejected industry’s claim that “nobody has shown any symptoms of lead poisoning”: “I cannot escape feeling that a hazard is perfectly clearly shown … here today”, and it could be called safe only “after very careful and prolonged and devoted study”. The authors comment: “This did not happen until the 1970s” (p. 52). - Dr Touart, who had treated the workers: “this ethyl gas is under suspicion and therefore should be withheld from public consumption until it is conclusively shown that it is not poisonous” (p. 52). - Haven Emerson (Columbia): using deaths as the indicator was unsound. Data on “functional and mental disabilities” of survivors were needed (p. 52). This is an early statement of endpoint choice. - Some public health scientists supported industry: the benefits were solid and direct, while evidence of public risk “was not available”. The authors: “This asymmetry between short-term economic benefits and long-term health hazards is another continuing problem” (p. 52). - Hayhurst at the meeting: 27 months of public use “should have sufficed to bring out some mishaps and poisonings”. The authors call these “two weak assumptions”: that statistics “collected for other purposes” could show the safety of a new technology, and that two years is enough to uncover new hazards. “These assumptions are still common in debates on current health hazards” (p. 52). - Private versus public views: Hayhurst had written privately to the PHS a week earlier sharing the concerns of PHS scientist Dr Thompson, who said “lead has no business in the human body … everyone agrees lead is an undesirable hazard and the only way to control it is to stop its use by the general public” (p. 53). Hayhurst continued: “Personally, I can quite agree with Dr Thompson’s wholesome point of view but, still, I am afraid human progress cannot go on under such restrictions … if we are to survive among the nations. Dr Thompson’s arguments might also be applied to the thousand and one other poisons” (p. 53). The authors read this as scientific judgement “influenced by political and economic factors”. The argument has two parts: national competitiveness and a slippery slope (if we restrict this, we must restrict everything). - Alice Hamilton, “the country’s foremost authority on lead”: “You may control conditions within a factory … but how can you control the whole country?” (p. 53). She later wrote that even the strictest factory conditions produced poisoning “sooner or later” (Hamilton 1925). - Frank Howard (first president of Ethyl) closed for industry: - “You have only one problem … is this a public health hazard?” Industry had others, such as ensuring cars and oil played a key role in national progress. - “Our continued development of motor fuels is essential in our civilisation”. TEL after a decade of research was an “apparent gift of God”. - “Because some animals die and some do not die in some experiments, shall we give this thing up entirely?” (p. 53). - The authors’ reading: “In a couple of rhetorical sentences he put the burden of proof back onto the public health scientists”. He also made them “appear to be reactionaries who were retarding human progress and technological innovation” (p. 53). - The meeting ended “after less than seven hours”. Ethyl announced a temporary suspension of sales pending a PHS “blue ribbon committee”. Hamilton expected the right result, especially with “a blaze of publicity turned on their deliberations” (p. 53).

3.4.2 Blue ribbon committee findings (pp. 53, 56) - The committee was under “great time pressure” and ran a seven-month study of 252 garage and filling-station attendants, chauffeurs and factory workers (p. 53). - It concluded: “at present there are no good grounds for prohibiting the use of ethyl gasoline … provided that its distribution and use are controlled by proper regulations” (p. 53). - Committee member Winslow’s recommendation to continue searching for alternatives “was omitted from the final committee report” (p. 53). - The report carried explicit caveats: data were incomplete, and widespread use might create “conditions … very different from those studied by us”. “Longer exposure may show that even such slight storage of lead … may lead eventually in susceptible individuals to recognizable lead poisoning or chronic degenerative disease”. “The committee feels this investigation must not be allowed to lapse” (p. 53). - The committee urged the PHS to seek funds from Congress for long-term prospective studies of consequences “not now foreseen”. “However, the PHS never undertook such research. For the next 40 years all studies of TEL were conducted and funded by the Ethyl Corporation and GM” (Markowitz and Rosner 2002; p. 56). - What exactly was conditional. In the committee’s formal conclusion the only proviso is “proper regulations” (p. 53). The long-term study appears as a caveat and a strong recommendation (pp. 53, 56), not as a stated condition. Table 3.1 nonetheless summarises the outcome as approval “only under careful monitoring and regulations, which do not take place” (p. 71). - 1926: the PHS recommended a maximum of 3 cc TEL per gallon. “Ethyl quickly agreed to comply, relieving the government of any pressure to introduce the regulations on lead in petrol that had been called for by the expert committee” (p. 56). This is voluntary compliance pre-empting binding regulation. - “For the next 35 years lead toxicity as a public health issue virtually disappeared from sight”. At peak TEL production “some 250 000 tonnes of lead were released into the air in the United States every year” (p. 56).

Panel 3.1 — Kovarik: the alcohol alternative (pp. 54–55)#

• 1907–08: USGS and US Navy ran more than 2 000 tests. Alcohol had “many advantages” in cleanliness. USGS later called it “a more ideal fuel than gasoline” with “better efficiency despite the high cost” (p. 54).
• GM’s energy-security motive is documented in an unpublished DuPont history (Wescott 1936). GM wanted “to fortify itself against the exhaustion or prohibitive cost of the gasoline supply”, which was expected in about 25 years. High-compression engines “could more advantageously be switched to [ethyl] alcohol” (p. 54).
• Midgley’s own advocacy:
• In an internal memo he called alcohol the “most direct route … for converting energy from its source, the sun, into a material that is suitable for a fuel” (p. 54).
• In October 1921 he drove a high-compression car on a 30% alcohol blend to an SAE meeting. “Alcohol has tremendous advantages and minor disadvantages”: clean burning, no carbon deposit, high compression without knocking, more horsepower (p. 54).
• Cellulose farm wastes could supply “enough alcohol to run our automotive equipment in the United States”, but it would cost two dollars per gallon. Oil shale would not work, and benzene from coal would meet only about 20% of need (Midgley 1921; p. 54).
• The reversal: once they had invested in TEL production, Midgley and Kettering “categorically denied the existence of alternatives”. NYT, 7 April 1925: “tetraethyl lead is the only material available which can bring about these [antiknock] results … unless a grave and inescapable hazard exists in the manufacture … its abandonment cannot be justified” (p. 54).
• Footnote 4 (Kovarik 2003) says TEL “crashed through the modest defenses of the American public health system of the 1920s not only through brute force of industry’s political influence over government but also due to the disorganized information resources available to public health advocates” (p. 54). This is information asymmetry about alternatives.
• A Commerce Department report, “published but not released” days before the conference, showed alternative anti-knock agents (mostly ethanol blends) “used routinely in two dozen other industrial nations” (pp. 54–55). Anyone familiar with Midgley’s 1921–22 papers “would see that by 1925 he was contradicting his own published research” (p. 55). Information about alternatives emerged at the conference only “in a few statements by public health scientists and hints in the media”, and “No record of any dissent exists” on this point (p. 55). (This slightly qualifies the chapter’s flat statement that “No ‘innovation’ other than TEL was discussed”, p. 52.)
• 1933: US military tests found Ethyl leaded petrol and 20% ethanol blends “almost exactly equivalent” in brake horsepower and compression. “This report was never published” (p. 55).
• Synthetic fuels and cartel: Fischer-Tropsch and Bergius coal-to-fuel processes were seen as serious competition to TEL. Standard Oil therefore entered a “full marriage” with I.G. Farben, dividing the chemical and fuel businesses “no matter how World War II progressed” (Davis 2007; p. 55).
• Institutional amnesia: “The wide variety of alternatives and substitutes known in the 1920s and 1930s were forgotten by the 1960s. Histories of the oil industry omitted any mention of alternatives” (p. 55). In 1974 Thomas Reed (MIT) began alcohol-fuel work “unaware of any other similar work before him” (p. 55).
• Defeating the competition (p. 55):
• A 1936 FTC restraining order told competitors to stop criticising Ethyl petrol, which “is entirely safe to the health of [motorists] and to the public in general … is not a narcotic in its effect, a poisonous dope” (US FTC 1936). A regulator thus enforced a safety claim and suppressed market criticism.
• “Fierce and unfair competition” from Ethyl led to a 1937 antitrust case. By then Ethyl was in 70% or more of US petrol (90% per Ethyl’s advertising) and in every major brand except Sunoco. Dealers who cut prices or sold alcohol or benzene blends were refused Ethyl fluid.
• A 1937 DOJ memo: Ethyl “exercised its dominant control … substantially to restrain competition”. Ethyl lost in 1938 and at the Supreme Court in 1940 and was ordered to supply any qualified customer (p. 55).
• Farm chemurgy / Agrol: a Depression-era populist movement promoted corn alcohol fuel. Agrol was sold at 2 000 stations at the same retail price as leaded petrol, but “it cost wholesalers and retailers an extra penny to handle it and this cut into their profit. By 1939, the Agrol plant had closed” (Hale 1934; Kovarik 2003, 2005; p. 55). The panel juxtaposes the two facts and implies that a small handling cost in the distribution chain helped sink the alternative. It does not show that this was the sole or main cause of closure.

Box 3.6 London buses (p. 56) - In 1919 the London General Omnibus Co. found alcohol–benzene blends about equal to petrol in mileage and favourable for “flexibility, absence of knocking and cleanliness”. The box notes in parentheses that “benzene is a dangerous carcinogen”. This is the only acknowledgement in the chapter that an alternative carried its own hazard. - The engineer said: “We are fast squandering the oil … it is to the fuels experimented with that we must turn for our salvation” (p. 56).

3.5 Lead contamination is “normal and safe” (pp. 56–58)#

• After 1926, Kehoe (University of Cincinnati) “was cultivated by the TEL industry as the dominant authority on lead” (p. 56). Kettering set up a Cincinnati laboratory with an initial USD 130 000 from Ethyl, DuPont and GM and asked Kehoe to direct it. Kehoe “later also became a corporate officer at GM and a consultant to DuPont” (p. 56).
• “The only source of funding for research came from industry sources”. The 1926 recommendation for publicly funded research “was not implemented” (p. 56).
• The control group error:
• Kehoe compared TEL workers with men in other jobs in the same plant, whom he designated “unexposed”. Finding lead in their excreta, he concluded that lead was “naturally present” and so “could not be very harmful”. “The mere presence of lead in workers … could not be an indicator of poisoning” (p. 56).
• Edsall, Henderson and others had attacked this in 1925, arguing that all Dayton workers were potentially exposed. The chapter records a concession: “Kehoe eventually came to see the merit in his critics’ assertions: clearly he had chosen the wrong control group” (p. 56). He then sampled a remote farming village outside Mexico City and again found lead (p. 56). This “‘natural’ lead levels” finding “became the nucleus of Kehoe’s position throughout his career” (pp. 56–57).
• The authors: he “assumed that general lead contamination was ‘normal’ and therefore ‘natural’ and harmless at those ‘low’ levels” (p. 57).
• Box 3.7 TEL and “treason” (p. 57): in 1942 Thurman Arnold told a Senate committee that without Ethyl’s leaded petrol the Nazis could not have fuelled their forces so efficiently. Chairman Harry Truman called the alliance with I.G. Farben “treason”. A German memo said TEL production was “entirely due to” American production plans and know-how supplied just before the war (Davis 2007). In the main text the authors describe the economics and politics of TEL as “plumbing new depths”. The links dated from the 1930s, when Ethyl formed the German company Ethyl Gemeinschaft and Standard Oil linked up with I.G. Farben (p. 57).
• “Industry control of both the economic and public agenda seemed complete by the 1950s”, but a challenge was “emerging from well outside the TEL community” (p. 57).
• Clair Patterson (Caltech geochemist):
• He had overturned the geological consensus on the age of the Earth (4.5 billion years against about 3 billion). His “extraordinary measures to avoid contamination” made his measurements far more accurate (p. 57).
• He treated reagent contamination “not [as] a nuisance but a clear indication of lead contamination from human activities” (p. 57). He studied deep-sea tuna, pre-Iron Age mummies and Greenland ice cores.
• His finding: modern body burdens “some 600 times higher than that of our pre-industrial ancestors” (p. 57). Table 3.1 (p. 71) instead says “100 times higher”, and p. 46 and p. 70 say “hundreds of times”. This is an internal inconsistency.
• Kehoe’s review of Patterson’s 1965 paper (“Contaminated and natural lead environments of man”, Archives of Environmental Health):
• He recommended publication so that Patterson could be demolished: “I should let the man, with his obvious faults, speak in such a way as to display these faults”.
• The inferences were “remarkably naïve”, from a man “woefully ignorant” and “so lacking in any concept of the depth of his ignorance”. “This bespeaks the brash young man … passionate supporter of a cause … It must be faced and demolished” (Kehoe 1965; p. 58).
• Vocabulary shift: Patterson argued that common does not mean harmless. “Normal” should be replaced by “typical”, and “natural” reserved for pre-contamination levels (p. 58). The paper also showed Kehoe’s “unexposed” subjects were contaminated, which “would dilute or hide risks of exposure”. The authors: “This dilution of risks by background contamination is now a much more common problem for public health” (p. 58).
• The response was a “fusillade of angry responses from orthodox toxicologists”, attacking his “hubris in stepping outside his field” and the journal’s editor. The authors’ generalisation: “shooting the messenger … pervades this and most other ‘Late Lessons’ stories, from John Snow and cholera … to those current scientists who publish warnings of hazards about climate change, genetically modified organisms and electromagnetic fields” (p. 58). The text dates Snow’s cholera work to “1864”. The passage is about how messengers are treated, not whether their warnings proved right. But it lists these issues together without distinguishing the state of the evidence in each.
• Two paradigms: Kehoe held that lead poisoning occurred “only at high doses with obvious signs of severe illness”. Patterson held that elevated levels in all humans were associated with “sometimes silent disturbances in body chemistry. Perhaps … everyone was poisoned to some extent” (p. 58).
• Retaliation:
• Ethyl visitors tried “to ‘buy me out through research support that would yield results favourable to their cause’” (Patterson 1992 letter to Needleman). He refused.
• His PHS contract was not renewed and his American Petroleum Institute contract was terminated. Caltech trustees visited his department chair “asking that he be fired” (p. 58).
• He predicted that future scientists would show Ethyl was poisoning people and that its operations would be shut down (p. 58).
• Later recognition: the Goldschmidt Medal, the National Academy of Sciences, an Antarctic peak and an asteroid (p. 58).
• The only source cited for the episode is Patterson’s own 1992 letter to co-author Needleman, and it is cited only for the “buy me out” quote (ref. p. 74). The contract losses and the trustees’ visit carry no citation. As with Henderson, the link to Ethyl is conveyed by sequence (“Following the meeting with Ethyl”).
• Box 3.8 Saul Bellow (p. 59): in The Dean’s December (1982), Professor Sam Beech is based on Patterson. “The genius of these evils was their ability to create zones of incomprehension. It was because they were so fully apparent that you couldn’t see them” (Bellow 1982). This is ubiquity as a cause of invisibility.

3.6 1966: US Congress asks awkward questions (pp. 58–60)#

• Senator Edmund Muskie’s Senate Subcommittee on Air and Water Pollution held hearings leading to the 1970 Clean Air Act. (The chapter misnames him “Edward Muskie”, p. 58.)
• Surgeon General William Stewart signalled concern about low-dose effects, “perhaps for the first time”. “Children and pregnant women constitute two of the most important of such groups. Some studies have suggested an association between lead exposure and the occurrence of mental retardation” (pp. 58–59).
• Kehoe as industry’s principal witness:
• Two years earlier he had said enough was known to allow increased TEL. “No other hygienic problem in the field of air pollution has been investigated so intensively, over such a prolonged period of time, and with such positive results” (p. 59).
• When pressed: “I have had a greater responsibility than any other persons in this country. The evidence at the present time is better than it has been at any time that this is not a present hazard” (p. 59).
• Asked whether a substitute would be desirable, he said lead is “an inevitable element” and “there is no way in which man has ever been able to escape the absorption of lead while living on this planet” (p. 59).
• This is the naturalisation argument, with authority resting on having done more research than anyone else.
• Patterson’s testimony:
• About 10 000 tonnes of lead are naturally cycled each year against “millions of tonnes” from industry and transport. The brain is the key target (p. 59).
• On regulator–industry collaboration: “It is not just a mistake for public health agencies to cooperate and collaborate with industries in investigating and deciding whether public health is endangered; it is a direct abrogation and violation of the duties and responsibilities of those public health organisations” (p. 59).
• Endpoints: industry had “traditionally measured the prevalence of lead toxicity by counting deaths”. Muskie asked whether low-level exposure might differ from classical poisoning (p. 59). Patterson: the organism “responds in a continuum … There is no abrupt change between a response and no response. Classical poisoning is just one extreme of a whole continuum of responses” (p. 59).
• The authors’ assessment: Muskie’s inquiry “marked the government’s shift away from complacency” and established that lead poisoning “could be an insidious, silent danger to the public”. The all-or-nothing notion was “replaced by degrees of disease spanning across a biological continuum of ‘effects’ to ‘adverse effects’”. The issue “still dominates current discussions about chemicals, radiation and other public health hazards, where early ‘effects’ are often dismissed as having no biological or ecological significance” (p. 60).

3.7 Lead in petrol poisons catalytic converters — so it’s got to go (p. 60)#

• GM sold Ethyl in 1962. In 1970 GM announced catalytic converters to meet the 1970 Clean Air Act, and said lead had to be phased out because it poisons the platinum catalyst. The authors: “Apparently, poisoning a technology was more important than poisoning people” (p. 60).
• Ethyl saw this as “a betrayal” and resolved to fight the environmental movement. It argued that TEL had “saved billions of dollars for the American economy and helped make the modern automobile possible”. It formed the “Ethyl Air Conservation Group” (“with unconscious irony”), staffed by Ethyl officials and Hunton and Williams lawyers and headed by an Ethyl board member (p. 60).
• EPA medical officers pushed for a separate health standard, fearing that “if a substitute for platinum were discovered sometime in the future, lead would return to fuel” (p. 60). They understood that the technology-compatibility rationale was not durable.
• 1973: aware of 200 000 tonnes a year from US car exhausts, the EPA issued a phase-down to 0.5 g/gal within five years (Schoenbrod 1980; p. 60).
• Industry exploited the oil crisis. The EPA estimated the oil penalty at 30 000 barrels a day. A full-page NYT advertisement on 2 December 1973 claimed that removing lead would be like “dumping one million barrels of oil a day” (p. 60). The gap is roughly 33-fold.
• On 6 December 1973 the EPA finalised the regulations. Ethyl and DuPont sued, arguing enormous cost, that no one had been poisoned by lead in air, and that low-dose effects “were not adverse health effects”. The first court set the rules aside as “arbitrary and capricious” (p. 60).
• On appeal the rules were upheld (Ethyl Corp. v EPA, 1976): “the regulatory action under this precautionary statute [the Clean Air Act] should precede, and hopefully prevent, the perceived harm”. The Administrator “may consider the cumulative impact of lead additives with other sources of human exposure to lead” (p. 60). Ethyl, PPG, DuPont, NALCO and the National Petroleum Refiners Association “then appealed to the Supreme Court, where they lost” (p. 60).

3.8 Public funding to study lead poisoning in children (pp. 60–61)#

• 1970: the Surgeon General called for identifying children with “undue” lead exposure, a statement that “avoided the loaded term ‘poisoning’”. “For the first time since 1925 significant research funds were allocated from Federal sources … The industrial monopoly on scientific data was drawing to an end” (p. 60).
• Needleman et al. 1979, described in the third person as “His seminal paper”. Needleman is presented as “one of the public health scientists who used the recently released public funds” (p. 60). This is the chapter’s main example of what public funding made possible.
• Higher lead was associated with greater negative impact on IQ.
• He shifted the paradigm from blood to “stocks of lead in the bones” by analysing “‘milk’ teeth from more than 2 000 infants” and correlating them with later intelligence and behaviour (pp. 60–61).
• A 5-point average IQ fall was “dismissed by industry as ‘small’”. That view “ignored the effect on very large groups of children who were at both ends of the normal distribution” (the numbers of severely handicapped children would double and of gifted children halve) (Bellinger and Bellinger 2006; fn 5 refers to Ch. 23 Fig. 23.1) (p. 61). This is the population-shift argument.
• “Industry responded to this dramatic observation with unprecedented opposition, resorting eventually to a character assassination of Needleman” (p. 61). No details are given of what the attacks consisted of or how they were resolved.
• Needleman followed up the 1943 Byers finding of chronic antisocial behaviour in children who had “recovered” from acute poisoning, confirming the link to adolescent antisocial behaviour (Needleman 1996). “Studies have further confirmed the link” (WHO 2010) (p. 61).
• Box 3.9 “It’s all about economics…” (p. 61):
• Needleman’s recollection of a late dinner “after … a liberal amount of red wine” with Jacobs from DuPont. DuPont’s economists had projected that gasoline consumption would level off and possibly decline, so the company “would not invest USD 100 million in research and development funds” in a safer additive.
• Needleman’s conclusion: “the entire debate about scientific studies, about the health risks for children, was merely a shadow play. The real decision had been made by DuPont’s economists. Their plan was clear: don’t budge on TEL and seek medical and environmental arguments to support the choice” (Needleman 2000; p. 61).
• This is a single-source, informal, recalled conversation. It is revealing but weak as evidence.

3.9 1977–1995: the phase down (p. 61)#

• EPA Air quality criteria for lead (Dec 1977): lead in air and dust was a significant exposure source, and brain damage could occur without acute symptoms. It was used to set an air lead standard (p. 61).
• Philadelphia air lead fell from 1.3–1.6 μg/m³ (1977) to 0.3–0.4 μg/m³ (1980), and similar trends appeared in most cities (p. 61).
• 1976–1980: lead used in petrol fell 50% and average US blood lead fell 37%. The EPA’s second criteria volume concluded, “contrary to the claims of the industry”, that the petrol–air lead relationship was causal: from 1975 to 1984 petrol lead fell 73% and composite maximum quarterly air lead fell 71% (USEPA 1986; p. 61).
• In the 1970s the toxic threshold was 60 μg/dl. Falling exposures “gradually allowed comparisons with children whose background blood lead levels were 1μg/dl or less”, revealing IQ effects below 10 μg/dl and “most of the cognitive impairment seeming to occur at blood lead levels as low as 5 μg/dl” (Lanphear et al. 2000; p. 61). This is an important methodological dynamic: reducing exposure made low-dose harm detectable.

Box 3.10 Blood lead declines (source WHO 2010; p. 62) - The 1976–1995 phase-out was associated with a more than 90% reduction in mean blood lead. - US children aged 1–5 at or above 10 μg/dl fell from 77.8% (late 1970s) to 4.4% (early 1990s). The mean was 1.9 μg/dl in 1999–2002. - Lead was also removed from can solder and new residential paint at the same time. The box states this fact but does not discuss it as a confounder. It matters for attributing the blood lead decline, and the IQ gain below, to petrol alone. The IQ gain is “associated with the decline in mean blood lead concentrations”, not with petrol specifically. - An estimated 5–6 point gain in mean population IQ is worth “between USD 100 billion and USD 300 billion in each birth cohort” (Grosse et al. 2002). - Similar declines occurred in Europe and in China, El Salvador, India, Mexico and Thailand. Unleaded petrol is about 99% of worldwide sales (p. 62).

Box 3.11 Continuous reductions in the “safe” level (p. 62) - CDC level of concern: 60 μg/dl (1960s), then 40 and 30 (1970s), 25 (1980s), and 10 (early 1990s), “the level that remains in place today” (Surkan et al. 2007). - A pooled seven-cohort analysis (Lanphear et al. 2005) found a rise from under 1 to 10 μg/dl associated with a 6 IQ-point decrement, “considerably greater” than from 10 to 20. This is a supralinear dose–response. - In 2004, 16% of children worldwide were above 10 μg/dl. In 2010 EFSA (miswritten “European Food Standards Agency”) withdrew its provisional tolerable weekly intake as inadequate to protect against IQ loss (p. 62).

3.10 The pros and cons of leaded petrol (pp. 62, 69)#

• “Leaded petrol was finally completely phased out in the US in 1995, seventy years since the ‘one day trial’” (p. 62).
• “The benefit of taking lead out of petrol exceeded the predictions of even the most convinced lead advocates” (p. 62). “Lead advocates” presumably means advocates of removing lead. The wording is ambiguous.
• Average US child blood lead: 13.7 μg/dl (1976), 3.2 (1991), 2.0 (2000) (p. 62).
• Health and other costs to taxpayers, health services and the economy “have been huge and have persisted for decades after the leaded petrol phase out, as contamination persists in soils and dusts” (Mielke 2010; cross-ref Box 23.1) (p. 62).
• The benefits side (p. 69): “Of course, leaded petrol also brought many benefits”: energy and fuel efficiency, thousands of jobs, and “much profit for the lead, oil and car industries”.
• The authors then argue: “These benefits could, however, have been attained by alternative uses of the economic capital involved. Indeed, a 10-year phase out of leaded petrol at any time since 1925 would have encouraged innovators to develop less hazardous and perhaps more efficient fuel additives and engine designs” (p. 69). Such innovations are “widely recognised as a useful defence against high oil prices and insecure oil supplies” (p. 69). This counterfactual is asserted, not demonstrated.

Panel 3.2 — Haigh: EU policymaking (pp. 63–64)#

• EU policy came mainly from the UK and Germany.
• UK 1971: the Chief Medical Officer advised that air lead should not rise. The government chose a three-stage reduction from 0.84 to 0.4 g/l by 1976, but “this phase down was then delayed” until an EC directive required it (p. 63).
• Germany 1971 chose a faster programme, implemented as planned: 0.4 g/l in 1972 and 0.15 g/l in 1976 (p. 63).
• “In both countries this initial action arose in response to scientific advice, without much public pressure”. The UK’s “more leisurely approach … was possibly linked to the fact that the largest European plant then manufacturing lead additives for petrol was in the United Kingdom” (p. 63). This is hedged.
• Germany’s unilateral move led the EC to set up two committees in 1971. They found “no immediate danger to public health” but judged it desirable to prevent increases. Cross-border trade justified EC-wide limits (p. 63).
• European Parliament, 10 November 1975: the second-stage 0.15 g/l limit “met with insurmountable opposition in the Committee because it would have required industry to make substantial investments and increased petrol consumption. Since these objections could not be refuted”, the Committee asked for postponement. It did approve the first-stage 0.4 g/l limit (p. 63).
• Council unanimity was “coloured by the existing German limit of 0.15 g/l”. Directive 78/611 set a 0.4 g/l cap and allowed national 0.15 g/l limits (p. 63).
• Haigh’s lesson: “in favourable circumstances a determined Member State can lead its peers despite considerable opposition and scientific uncertainty”, achieving higher standards faster (p. 63).
• UK 1981: after the Lawther report (1980) and WOPLIP (1979), the government set 0.15 g/l, “the lowest level that could be required under the Directive”, but “did not propose lead-free petrol”. The health and environment ministries “were defeated on the second point by transport, energy and the treasury” (p. 63). This is intra-governmental conflict.
• April 1983: the Royal Commission on Environmental Pollution recommended lead-free petrol for all new cars, and “the Government immediately accepted”.
• In between there was “an extraordinary public campaign” by CLEAR, “supported by a millionaire”, providing lobbying and publicity for the science.
• The RCEP might have taken up lead only because of the campaign, though its Chairman (Southwood) disagreed.
• “What can be said with some certainty is that the government only endorsed the Commission’s conclusions so quickly (within half an hour of publication) because of the campaign and because of an imminent general election” (p. 63).
• The EC’s reaction to the UK’s April 1983 approach was negative. Then German concern about forest die-back meant that NOx reductions needed catalytic converters, which meant removing lead. Germany, the Netherlands and Denmark backed the UK, and Directive 85/210 was adopted (p. 64).
• “Public conscience was excited by two quite different issues: public health and death of forests. It is pure chance that they came together at the same time and if either had been missing it is quite possible that the directive would not have been agreed, or not agreed so quickly.” Haigh adds: “what would have happened if someone had invented a lead tolerant catalytic converter?” (p. 64).
• On science and publics: science informs public opinion, “a major determinant of policy”, but “science itself does not always reach the public at a specific point in time when a specific decision is called for”. There is no single “European public”, only national publics, so policy debates differ between countries (p. 64).
• Figure 3.1 (p. 64; reprinted from Haigh 1998), captioned “Key events which help to explain reduction in lead levels in petrol in Germany, the United Kingdom and the EU”: an unlabelled stepped line of petrol lead limits falling from 0.84 through 0.64, 0.55, 0.50, 0.45 and 0.40 to 0.15 g/l over 1970–1990. The values match the UK sequence in the text. It is aligned with five event rows:
• Scientific advice: CMO, EC report, WOPLIP (misspelt “WOBLIP”), Lawther, RCEP.
• Force majeure: the Yom Kippur War / oil crisis around 1973.
• International activity: the German 0.4 and 0.15 levels, EC directive proposed and agreed, the German “change of heart on pollution”, and the EC decision on lead-free petrol.
• Pressure groups: Conservation Society, CALIP, CLEAR, BEUC/EER.
• Key UK decisions: phase reduction to 0.4 g/l; reduction to 0.15 g/l with “lead-free ruled out”; policy of lead-free with a request to the EC to amend the directive.

Panel 3.3 — von Storch et al.: the German experience (pp. 65–66)#

• Environment featured strongly in early-1970s German politics. Germany was the first European country to restrict lead: 0.4 g/l from 1972 (down from 0.6) and 0.15 g/l from 1976 (p. 65).
• According to “a preliminary analysis of newspaper coverage”, health dangers entered the German press in the 1960s, while British coverage “focused on urban smog” (p. 65). This is different agenda-setting in national media.
• Unleaded petrol (0.013 g/l) was introduced in October 1984. Germany could not ban leaded petrol because EC trade rules forbade it, so it used tax incentives (1984), mandatory availability at all stations (1985), and enhanced incentives making unleaded cheaper (1986) (p. 65).
• EC law: Directive 85/210 required unleaded to be available by October 1989 and asked states to adopt 0.15 g/l voluntarily. Directive 87/416 allowed Member States to ban leaded 92-octane (p. 65).
• By 1995 unleaded had more than 80% of the market in Germany, Sweden, Finland, Denmark, the Netherlands and Austria, but under 30% in France, Greece and Portugal. The key drivers were higher leaded prices and “lead-averse catalysts”, plus informing the public that unleaded was safe for non-catalyst cars (Löfgren and Hammar 2000; p. 65).
• European road lead emissions: 31 000 t (1955), 119 000 t (1975), 62 000 t (1985), 42 000 t (1990), 19 500 t (1995) (p. 65). The 1975 peak of 119 000 t conflicts with the chapter’s “about 200 000 tonnes … in both the US and Europe” (p. 46).
• “Favourable terms of competition were experienced by producers of cars with high technical standards, who had already gathered experience with catalyst systems on the US market” (Hagner 2000; p. 65). Regulation advantaged prepared firms.
• Modelling of blood lead (Figure 3.2, p. 66): the Münster and Germany models give three scenarios. 1. No regulation (0.6 g/l continued): mean PbB climbs to about 170–200 μg/L (Münster) or stays around 115–145 μg/L (Germany model). 2. No unleaded introduced in 1985: PbB plateaus around 75 μg/L (Münster) and about 55 μg/L (Germany). 3. Regulation already from 1961: lower curves.
• Reconstructions follow the upper curves to the early 1970s and then fall to about 40–45 μg/L by the mid-1990s.
• Units are μg/L, so 100 μg/L = 10 μg/dl. The no-regulation Münster curve (about 170–200 μg/L) is about 17–20 μg/dl. The reconstructions end at about 4–4.5 μg/dl.
• The panel’s conclusion: “In the case of no or delayed regulations, the model estimates that PbB levels well beyond the critical level would have emerged. Thus, the regulation instituted in Germany since the 1970s has reduced health hazards significantly” (p. 65). This is a model-based counterfactual on blood lead, not health outcome data. The “critical level” is not defined.
• Costs: “The macroeconomic costs of the regulation seem to have been insignificant in spite of concerns that they would be substantial” (Hagner 2000). “The case of leaded petrol demonstrated the limited utility of purportedly objective cost-benefit analyses, as the costs claimed at the time of the regulations turned out to be significantly biased, due to the vested interests that supported the analyses” (p. 65). The “significantly biased” claim comes with no figures and no citation beyond Hagner (2000).
• Closing: success “should not downplay the consequences of the introduction of tetraethyl lead”, since “alternatives were known and available already in the 1920s and 1930s” (Kitman 2000). Heavy metals accumulate in soil “which will remain for centuries”. “The strategy of protecting the environment from persistent substances must be based on continuous assessment and precautionary principles” (p. 66).

Panel 3.4 — Millstone: the UK experience (pp. 67–68)#

• 1930 UK Departmental Committee on Ethyl Petrol: advice came from Kehoe and from Surgeon General Cumming. Cumming “had moved from initial concern to the enthusiastic promotion of TEL writing dozens of letters touting Ethyl leaded petrol to public health leaders around the world”. Cumming reported to Treasury Secretary Andrew Mellon, “whose Gulf Oil Co. had exclusive contracts to distribute Ethyl petrol in the south-eastern US”, which “may have had something to do with his enthusiasm” (p. 67). A regulator becomes a promoter, and a possible conflict of interest in the chain of command (hedged).
• The committee concluded that widespread use would not raise street lead enough “to constitute a risk even to the health of that part of the population which is most exposed — namely, police officers on traffic control duty and drivers”. Given these assurances, TEL came into use in the UK and Europe (p. 67). This is how the US decision was exported through expert assurance.
• Late 1970s: evidence of high UK exposure and low-dose toxicity led to the Lawther committee. Lawther (1980): “We have not been able to come to clear conclusions concerning the effects of small amounts of lead on the intelligence, behaviour and performance of children”. This was “subsequently repudiated by several members of the committee for having been overly timid” and criticised by others (Rutter 1983; Bryce-Smith and Stephens 1980). Lawther nonetheless advised progressive reduction “without explaining why” (p. 67).
• The government “tried to represent the report as if it had proved that children’s blood lead levels were entirely harmless”. BP and Associated Octel “continued to downplay the toxicity”, and the government resisted reductions (p. 67).
• “In general, scientists adopted a very cautious approach.” The MRC (1988) said the associations were “consistent with the hypothesis that low-level lead exposure has a small negative effect”, but “it is not possible to conclude that exposure to lead at current urban levels is definitely harmful” (p. 67).
• RCEP (1983): “We are not aware of any other toxin which is so widely distributed in human and animal populations and which is also so universally present at levels that exceed even one tenth of that at which clinical signs and symptoms occur” (p. 67).
• Action levels:
• Lawther recommended investigating children above 35 μg/dl, although evidence already pointed lower (Chisholm 1976; Needleman 1979). It recommended no screening programme. Its sensible recommendation on detecting lead paint in inner cities “30 years later has not yet been properly implemented” (p. 67).
• The 35 μg/dl figure merely restated an existing EEC-wide agreement. The Department of Health moved to 30 μg/dl in 1981 (p. 68).
• In 1982 the UK set an “action level” of 25 μg/dl “by reference to the results of a blood lead survey rather than toxicological considerations”. Because most people were already below it, it “necessitated no further remedial action. This exemplifies the British government’s practice of not setting lead targets until they had already been met” (p. 68).
• Through the 1980s, evidence of neurotoxicity at lower levels grew (US, Greece, Australia). The government’s response to CLEAR was “the classic tactic of establishing yet another investigative committee” under the MRC (p. 68).
• Question framing (Millstone’s analysis):
• The government asked the MRC panel: “does the evidence on childhood neurotoxicity prove that levels of lead in British children are doing them obvious harm?”
• This “set a particularly high evidential bar: indicative evidence short of proof would be insufficient”. It did not ask “in which physiological system(s), and at which lowest level of exposure, are adverse effects detectable?”
• “If it had asked a question of that sort, a rather different answer would have been obtained. The government proposed only to act in the face of compelling evidence rather than, for example, the balance of probabilities” (p. 68).
• MRC 1984 “sat resolutely on the fence”, emphasising methodological limits and saying any effect was small. MRC 1988 accepted “a small negative effect” and concluded “it would be prudent to continue to reduce” exposure. Millstone reads this as tacit acknowledgement that exposures were unacceptably high, “couched in language designed not to provoke public anxiety” (p. 68).
• Internal tension. On p. 67 Millstone quotes MRC 1988 as saying the associations are “consistent with the hypothesis that low-level lead exposure has a small negative effect”, and that “it is not possible to conclude” current urban levels are “definitely harmful”. On p. 68 he says the same report “accepted that ‘low level lead exposure has a small negative effect’”, apparently quoting the words that p. 67 presents as a hypothesis. The p. 68 characterisation may overstate how far the MRC went.
• Millstone also says the scientists’ cautious approach was “marginally modified” after the RCEP’s 1983 statement (p. 67).
• 1987: a preferential tax on unleaded was adopted “to facilitate the use of catalytic converters … rather than in response to evidence of lead’s neurotoxicity” (p. 68).
• “Curiously, official efforts to monitor childhood blood lead levels in British children then came to an end”, so UK benefits are “only fragmentarily documented” (p. 68). Monitoring was abandoned once action was taken, which prevents evaluation.
• The UK was “one of the last industrialised countries to embrace unleaded petrol”. Attributing the slowness to Associated Octel’s UK location is “difficult to establish” (hedged). “Nonetheless”, in summer 2010 two former Octel executives “were convicted of having bribed government officials in Indonesia and Iraq to continue allowing the use of tetra ethyl lead” (Leigh et al. 2010, Guardian) (p. 68).

3.11 European reflections (p. 69)#

• Other countries’ campaigns “went through similar phases and arguments” (p. 69).
• Legacy lead from petrol, old mines and lead shot still threatens some European children (EFSA 2010) and wildlife (Mateo et al. 2007; Rodriguez-Estival et al. 2012) (p. 69).
• Box 3.12 E-waste (p. 69): much electronic waste is shipped to low-income countries, where formal and informal workers, many of them children, recover metals. Elevated lead in dust and blood has been found (Huo et al. 2007, Guiyu). The box repeats verbatim two sentences from Box 3.11 (the 16% figure and EFSA 2010), which looks like a production error.

3.12 Some late but contemporary lessons (pp. 69–71)#

The lessons fall into two groups, science and “the influence of society on the science”. Epidemiological arguments are deferred to Chapter 26 (p. 69). They are listed in full in the lessons section below.

Table 3.1 Early warnings and actions (p. 71)#

The chronology runs from the 2nd century BC (Nicander) to 2013. Entries not in the main text include: - 1695: the Count of Württemberg bans lead in wine after Gockel’s Ulm study. - 1920: leaded paint banned in Australia. - 1983: an EC lead-isotope study in northern Italy “demonstrates that petrol additives cause substantial human exposures”.

The table also introduces characterisations harder than the text: - “1921–1923 Ethanol-based alternative additives are considered by Du Pont and GM but rejected as less profitable than TEL”. The text says “likely to be much less profitable” (p. 50). - “1925 … approval by an expert committee but only under careful monitoring and regulations, which do not take place”.

Discrepancies with the text: - “1923–1924 Deaths of TEL workers lead to its temporary suspension”. The text dates Ethyl’s suspension announcement to the May 1925 conference. - “1930–1960s Kehoe and the TEL industry dominate the research field for next 50–60 years” (sic). - “1965 Patterson reports … 100 times higher” (the text says 600 times, p. 57). - “1984 Germany introduces low-lead petrol” (Panel 3.3 says unleaded, p. 65). - “1920 Leaded paint is banned in Australia and later in Europe”. Box 3.4 dates the European and Australian bans to “between 1909 and the 1930s” (p. 49). - “1995–2000 Virtually all western Europe only uses lead-free petrol”. Panel 3.3 gives unleaded under 30% of the market in France, Greece and Portugal in 1995 (p. 65), and the chapter summary dates the European phase-out to 1998–2005 (p. 46).

Case timeline#

Lags - From the first specific warning (Clark, October 1922) to the US phase-down rule (December 1973): about 51 years. To complete US on-road phase-out (1995/96): about 73 years. - The chapter’s own figure: “50–70 years” from Henderson’s 1925 warning to phase-out in the US and Europe (p. 47). - From the committee’s call for long-term public studies (1926) to significant federal research funds (1970): 44 years (p. 60). - From the Patterson paradigm challenge (1965) to the US rule (1973): 8 years. - UK: from the 1930 assurance to the 1971 CMO advice took 41 years. Formal lead-free policy followed in 1983 and a tax incentive in 1987.

What was known when - By 1925 it was known that lead is a cumulative, insidious poison (Cumming 1922; Oliver 1911). Children’s poisoning from paint was well documented (Box 3.4). Organic lead killed and maddened workers (Box 3.5). Clark predicted that lead oxide dust would linger at street level (Clark 1922), but this was a prediction: in 1926 the committee still could not say “definitely whether exposure to lead dust increases in garages” (p. 53). Alternative anti-knock agents and fuels existed (Panel 3.1). - Not known in 1925: the actual public exposure from exhaust, and whether chronic low-level exposure harmed the public, especially children’s development. The committee acknowledged these data gaps (p. 53). - The dispute in 1925 was therefore not over whether lead was toxic. It was over who bore the burden of proof for a new diffuse exposure, and what evidence would suffice (pp. 52–53). - The decisive low-dose child evidence (Needleman 1979; Lanphear 2000/2005) came decades later. The US phase-down (1973) preceded Needleman 1979. It followed GM’s 1970 catalyst announcement, but the EPA framed its final rule as “to protect health”, EPA medical officers pushed for a health standard separate from catalyst compatibility, and the rule was upheld under a “precautionary statute” (p. 60).

Harms and costs (as reported) - Workers: at least 11 deaths and hundreds poisoned in 1923–24 (p. 51). - Emissions: about 200 000 t a year of airborne lead in the US at the mid-1970s peak (pp. 46, 60), or 250 000 t (p. 56). In Europe, 119 000 t in 1975 (p. 65), or “about 200 000” (p. 46). - Population blood lead: US child mean 13.7 μg/dl in 1976, with 77.8% of children aged 1–5 at or above 10 μg/dl in the late 1970s (p. 62). 16% of children worldwide were above 10 μg/dl in 2004 (p. 62). - Economic: a 5–6 point IQ gain from the decline, worth USD 100–300 billion per US birth cohort (p. 62). Continuing remediation and litigation costs from legacy lead (pp. 49, 62). Cross-reference to Chapter 23 costs of inaction (p. 62). - Legacy: soil contamination “for centuries” (p. 66). E-waste exposures in low-income countries (p. 69).

The authors’ own lessons and conclusions#

3.12.1 General lessons from the science (pp. 69–70)#

1. Sensitive subgroups. Early evidence came from “fit, adult, usually male workers” and was seen as irrelevant to lower public exposures. But the public includes more sensitive subgroups (children, foetuses, the elderly, pregnant women and so on). They are exposed around the clock through multiple routes. “Great care must therefore be taken in assuming that evidence from highly exposed occupational groups, or from low exposures to average populations, is not relevant to sensitive public groups” (p. 69).
2. Acute endpoints mislead. Reliance on mortality or acute poisoning is “a poor guide to long-term (chronic) effects on morbidity such as neurological or reproductive damage” (pp. 69–70).
3. Assertion in place of demonstration. Key assumptions “were confidently asserted rather than demonstrated”, as in Midgley’s “no actual experimental data”. Assuming “‘no evidence of harm’ is the same as ‘evidence of no harm’” is “still a common mistake in public health” (p. 70).
4. The unsupported “balance” assumption. It was assumed that excretion balanced intake, so lead would not accumulate. This “was not supported by actual evidence” (p. 70). (The chapter body never discusses this assumption or says who held it. The lesson appears only here.)
5. Contaminated controls. “When ‘unexposed’ control groups are also contaminated then true risks will be underestimated” (p. 70).
6. An underpowered first study of consumer risk was not followed by the recommended publicly funded long-term monitoring (p. 70).
7. Animal evidence was discounted. Animal studies at environmentally relevant concentrations “was frequently ignored or regarded as irrelevant for humans” (p. 70). (The chapter body cites no animal study that documented adverse effects. The only animal evidence in the body is the Bureau of Mines study, which reported no public risk (p. 51), and Howard’s “some animals die and some do not die” (p. 53), which implies some adverse animal results were discounted.)
8. Manufactured doubt. Debates “sometimes focusing on diversionary details, or based on the potential for exploiting or even manufacturing scientific doubt” kept harms looking “unproven”. “It is often more convenient for a hazardous industry to debate the science than to discuss options for reducing hazards” (p. 70).
9. No threshold between effect and adverse effect. The assumed threshold is “still a dominant assumption in conventional toxicology”. There is “usually a biological continuum and not a discrete change”, so “action to avoid significant biological ‘effects’ will often be needed if we are to prevent, as opposed to merely observe, ‘adverse effects’” (p. 70).

3.12.2 The influence of society on science (pp. 70–71)#

1. Industry monopoly on evidence. For decades “virtually no independent research” was done. It took scientists from outside, “such as Patterson in the 1960s and Needleman in the 1970s”, to show that typical body burdens were not “normal” (p. 70).
2. Recommendation: provide “sufficient incentives and funds for independent long-term prospective monitoring of potential health hazards when new technologies are introduced” (p. 70).
3. Unequal contest. Established, specific benefits accruing “to particular and powerful minorities” were set against unproven, general, future threats. “This was an unequal contest, which even influenced many public health specialists”, who let their appreciation of “the gift from God” override their concerns (p. 70). (The lesson attributes the phrase to “the car industry”. On p. 53 it is Ethyl’s president, Frank Howard, who says it.)
4. Advocacy role for scientists. Public health is well served when scientists who discover hazards, “especially when funded by the public, play an active role in disseminating both their results and their implications for precautionary or preventive action” (Hamilton, Henderson, “Craig” [Clair] Patterson, Needleman) (p. 70).
5. Protect early warners. Each wave of early-warning scientists (Henderson, Byers, Patterson, Needleman) had funding withdrawn, jobs threatened or characters assassinated. “Such scientists need more support from society via recognition for their work, help with their defence and legal protection against discrimination” (cross-ref Ch. 24) (pp. 70–71). Caveats: the chapter body gives no account of Byers being harassed; he appears only as the author of the 1943 finding (p. 61). This lesson dates Needleman to “the 1980s”, while society lesson 1 has “the 1970s”.
6. Open the record. “The concrete record of decision-making by industries, scientists and governments need to be made publicly available if history is to stand a reasonable chance of being understood.” This “usually only occurs many years after the relevant events and then only via legal cases for compensation” (p. 71).

Lessons stated elsewhere in the text or panels (not in 3.12)#

• Authoritative assertions are a recurring pattern across Late lessons cases (asbestos, CFCs, BSE) (p. 50).
• Censoring of hazard words (“ethyl” for lead; “tumour” for cancer) (p. 50).
• Weak safety inferences drawn from statistics collected for other purposes and from short observation periods (p. 52).
• Asymmetry of short-term economic benefits against long-term health hazards (p. 52).
• Dilution of risk by background contamination is a growing problem (p. 58).
• “Shooting the messenger” is pervasive (pp. 48, 58).
• Early “effects” are dismissed as insignificant in current chemical and radiation debates (p. 60).
• Counterfactual innovation claim: a 10-year phase-out would have spurred safer innovation (p. 69).
• Haigh (panel): a determined Member State can lead its peers (p. 63). Outcomes depended on “pure chance” coincidence of issues (p. 64). Science does not always reach the public when decisions are made (p. 64).
• Von Storch et al. (panel): limited utility of “purportedly objective” CBA because claimed costs were biased by vested interests (p. 65). Persistent substances require “continuous assessment and precautionary principles” (p. 66).
• Millstone (panel): the question posed to experts sets the evidential bar (p. 68). Targets set only once met (p. 68). Committees used as a delaying tactic (p. 68). Monitoring discontinued after action (p. 68).

Evidence-derived lessons versus recommendations and advocacy#

• Derived from the case evidence: science lessons 1–3, 5, 6 and 9, and society lesson 1. These are well supported by the narrative and by the later science (Boxes 3.10–3.11).
• Interpretive generalisations with thinner support in this chapter:
• science lesson 4 (the “balance” assumption is never discussed in the body);
• science lesson 7 (no animal study showing adverse effects is cited);
• science lesson 8 (manufactured doubt), which is hedged in the lesson itself (“sometimes”, “potential for exploiting or even manufacturing”). Its main support here is the Kehoe episodes and Needleman’s recalled Box 3.9 conversation (“seek medical and environmental arguments to support the choice”, p. 61).
• society lesson 5 also names Byers, whose harassment the chapter does not document.
• Recommendations or advocacy: society lessons 2, 4, 5 and 6; the 10-year phase-out counterfactual (p. 69); von Storch’s precautionary strategy (p. 66); and the extension to mobile phones (p. 50 fn 3) and to climate/GMO/EMF messengers (p. 58).

Mechanisms and dynamics#

1. How the warnings arose. Warnings came from several independent sources: - Deep historical knowledge: centuries of occupational and consumer poisoning (pp. 48–49). - Mechanistic reasoning about new exposure pathways (Clark on street dust, p. 50; Cumming on cumulative intake, p. 50). - Dramatic occupational sentinel events (Box 3.5, p. 51). - Clinicians who treated victims (Touart, p. 52). - Much later, outsider measurement science (Patterson’s contamination-free geochemistry, p. 57) and new biomarkers (Needleman’s teeth, pp. 60–61).

The 1920s warnings were strong on hazard and plausibility but lacked exposure–outcome data for the public. That gap was the lever for delay.

2. How warnings were contested. The chapter documents a repertoire: - Authoritative assertions without data (Midgley 1922; Standard Oil; p. 50–51). - Commissioned studies with sponsor control (Bureau of Mines: word substitution, draft approval; p. 50). - Reassurance at the moment of crisis: the Bureau report came out the day after the fifth death (p. 51). The chapter notes the timing but does not claim it was deliberately arranged. - Conflicted experts presenting as independent (Hayhurst; p. 52). - Blaming victims: workers’ “carelessness” (pp. 51–52). - Distinguishing exposures: the factory is not the street (p. 52). - Invoking progress and necessity: “gift of God”, “essential in our civilisation”, “survive among the nations” (p. 53). - Reversing the burden of proof: “shall we give this thing up entirely?” (p. 53). - Naturalising contamination: “normal”/”natural” and “no way … to escape the absorption of lead” (pp. 56–59). - Personal attacks on critics (Kehoe 1965; p. 58). - Inflated cost claims (one million barrels a day; p. 60). - Litigation (p. 60). - Reframing low-dose effects as not “adverse” (p. 60).

3. Burden and standard of proof. This is the analytical centre of the case. - In 1925 critics proposed a strict pre-market standard. Henderson demanded “absolute proof … that no danger to the public would be involved” (p. 51). Touart wanted withholding “until it is conclusively shown that it is not poisonous” (p. 52). - Industry successfully placed the burden on critics (Howard, p. 53). - The committee’s “no good grounds for prohibiting … provided that … controlled by proper regulations” (p. 53) was a conditional approval. Its formal condition (regulation) and its strongly urged safeguard (long-term publicly funded study) were never implemented (p. 56; Table 3.1 treats both as conditions, p. 71). - The balance shifted only under a statute that a court read as “precautionary”, permitting action to “precede, and hopefully prevent” harm and to consider cumulative exposure (p. 60). - In the UK, the government’s question to the MRC (“prove … obvious harm”) set the standard at proof rather than “the balance of probabilities” (p. 68). - The recurring point: whoever frames the question fixes the evidence threshold, and hence the outcome.

4. Control of knowledge production and funding. - For about 40 years all TEL research was industry-funded (p. 56). The dominant expert ran an industry-funded lab and held industry posts (p. 56). - The government’s own research agency accepted gagging terms (p. 50). The public health service never funded the follow-up its own committee demanded (p. 56). - The knowledge base was not just biased but narrowed. Endpoints (death, acute signs), comparators (contaminated “controls”) and concepts (“normal”) were set within the industry-funded paradigm (pp. 56–57, 59). - Independent public funding from 1970 is presented as ending “the industrial monopoly on scientific data” (p. 60).

5. Mental models of proponents and experts. - Midgley and Kettering: engineering confidence that dispersal would dilute exposure to the undetectable (p. 50), and an instrumental relationship to alternatives, which were championed until investment committed them to TEL (p. 54). - Howard: TEL as a providential solution. Health was the public health scientists’ “one problem”, while industry carried the nation’s progress (p. 53). - Hayhurst: private agreement with the critics, overridden by a belief that restriction was incompatible with progress and national survival, and by a slippery-slope fear (p. 53). - Kehoe: a threshold model of toxicity (poisoning only with overt illness, p. 58) and “natural” as a synonym for safe (p. 57). The intake–excretion “balance” assumption (lesson 4, p. 70) is not attributed to anyone in the chapter. Linking it to Kehoe is this note’s inference. His authority was grounded in the volume of his own research (“greater responsibility than any other persons”, p. 59). He treated outsiders as ignorant trespassers (p. 58). The chapter shows how a coherent, confident, well-funded paradigm can persist for decades when its core assumptions (clean controls, thresholds) are never tested against an independent baseline. - Cumming: moved from concern to promotion (p. 67). - Critics: Hamilton’s point was about controllability. Hazards can be managed inside a factory but not across “the whole country” (p. 53). Patterson’s point was the continuum model (p. 59).

6. Framing and language. - “Ethyl” instead of lead (p. 50). - “Gift of God” and “industrial progress” (p. 53). - “Normal/natural”, reframed by Patterson as “typical/natural” (p. 58). - “Undue” exposure rather than “poisoning” (p. 60). - “Prudent” as MRC language “designed not to provoke public anxiety” (p. 68). - “Ethyl Air Conservation Group” (p. 60). - Media labels from the other side: “Mad gas”, “house of butterflies” (p. 51). - The FTC’s enforced description of Ethyl as “entirely safe” (p. 55). - Language shaped what was visible. Bellow’s “zones of incomprehension” makes the same point: harms so ubiquitous they cannot be seen (p. 59).

7. Asymmetry of benefits and harms; distribution. - The benefits were concrete, immediate, measurable and “largely” concentrated. They went to “particular and powerful minorities”: the lead, oil and car industries (pp. 69, 70). The chapter also lists more diffuse benefits, namely vehicle energy and fuel efficiency and “thousands of jobs” (p. 69). - The harms were diffuse, delayed, statistical and fell hardest on children, foetuses, poorer neighbourhoods and later low-income countries (pp. 47, 49, 69, 70). - Workers bore the earliest and most acute harms and were blamed for them (p. 51). - Remediation costs fall on taxpayers and health services (p. 62). - The small individual-level IQ effect has large population consequences at the distribution tails (p. 61). Diffuse harm is systematically under-weighted when judged at the individual level.

8. Lock-in and path dependence. - Henderson predicted the lock-in itself: “nearly universal use” before society “awakens” (p. 47). - Market lock-in: Ethyl in 70–90% of US petrol by 1937; control of distribution; refusal to supply dealers who sold alternatives; the FTC’s gag on criticism; a Standard–Farben market-division pact (p. 55). - The Agrol episode suggests how a one-penny handling cost in the distribution chain could undercut an alternative sold at retail price parity (p. 55). The panel implies this link but does not establish it as the cause of closure. - Cognitive and institutional lock-in: alternatives “forgotten by the 1960s”, company histories omitted them, and a 1974 researcher was unaware of prior work (p. 55). - The installed vehicle fleet slowed transition, since air lead fell partly with “the gradual retirement of old cars” (p. 61).

9. Alternatives and substitutes. - The chapter’s position is that a cleaner, “equally effective” alternative existed (p. 46). Its evidence (Panel 3.1) shows technical feasibility and performance equivalence (1933 tests). - But Midgley put the cost of alcohol from farm cellulose wastes, at the scale needed to run all US automotive equipment, at USD 2/gal (p. 54), and USGS noted alcohol’s “high cost” (p. 54). The figure is for a national alcohol fuel supply, not for the 20% blends tested in 1933, so it is not a like-for-like comparison. The chapter does not systematically compare costs or supply at scale. - It names profit, not technical infeasibility, as the reason for rejection (p. 50; Table 3.1). - The only acknowledgement of substitute risk is benzene’s carcinogenicity (p. 56). The chapter does not discuss the hazards of the additives that replaced lead after phase-out.

10. Technological co-drivers of action. - In all three jurisdictions, the move to fully unleaded petrol coincided with catalytic converter compatibility: GM in 1970 (p. 60), Germany’s forest die-back and NOx concerns in 1983–85 (p. 64), and the UK tax differential in 1987 (p. 68). Market uptake in Europe was driven by catalyst adoption and price (p. 65). - The earlier phase-downs did not depend on catalysts. The UK’s 1971 decision and Germany’s 1972/1976 limits “arose in response to scientific advice, without much public pressure” (p. 63). The US EPA’s 1973 rule was framed “to protect health” and upheld under a “precautionary statute” (p. 60). - EPA staff saw catalyst dependence as fragile and pushed for a health-based standard in case a lead-tolerant catalyst appeared (p. 60). Haigh posed the same counterfactual (p. 64). - Health evidence produced partial reductions on its own. For elimination, an independent technological or policy driver, aligned partly by chance, was decisive in the accounts given.

11. Institutional behaviour of regulators and governments. - The regulator acted as a research contractor under industry terms (Bureau of Mines, p. 50) and as a promoter (Cumming, p. 67). - Voluntary compliance pre-empted regulation (p. 56). Committee recommendations were edited (Winslow’s dropped, p. 53) or unimplemented (p. 56). - In the UK: health and environment ministries were overridden by transport, energy and treasury (p. 63). An action level was set by reference to a blood lead survey rather than toxicology, so that it “necessitated no further remedial action” (Millstone’s account, p. 68). “Yet another investigative committee” was used as a response to pressure (p. 68). Monitoring stopped after the policy change (p. 68). - A domestic producer possibly slowed national action (p. 63, hedged). The EP committee judged industry investment objections impossible to refute (p. 63). - Against these: the courts (p. 60), the legislature (Muskie, pp. 58–60) and the RCEP (p. 63) acted as corrective forums.

12. Law, trade and markets. - The antitrust litigation (1937–40) addressed market conduct, not health (p. 55). - EC single-market rules prevented Germany from banning leaded petrol, so it used tax instruments (p. 65). EC cross-border trade in petrol justified harmonised limits (p. 63). - Council unanimity let a frontrunner keep its stricter rules: Directive 78/611 had to allow national 0.15 g/l limits, though its main provision was a 0.4 g/l upper limit (p. 63). - International licensing spread TEL to Germany (p. 57). Two former senior executives of a UK producer (Octel) were reported convicted in 2010 of bribing officials abroad to keep TEL in use (p. 68). This suggests that phase-out in rich markets displaced sales to weaker jurisdictions. The chapter does not state this generalisation; it is this note’s inference from one case.

13. Media, publics and campaigns. - The 1924 press coverage was followed by a sales ban (“New York State”, p. 51) and added to the pressure on the PHS and industry (p. 51). The chapter attributes the 1925 conference itself to “requests for action from public health experts” (p. 52). NYT reporters uncovered more cases than the companies disclosed (p. 51). - German press coverage from the 1960s preceded action (p. 65). - The UK’s CLEAR campaign (with private money) and an election produced a half-hour policy reversal (p. 63). - Hamilton hoped publicity would discipline the committee (p. 53). It did not. - Haigh’s observation: science reaches publics unevenly and not always in time for decisions (p. 64).

14. Energy security as a dual-use argument. - Energy security was mobilised by both sides. Alcohol was promoted as a hedge against oil exhaustion in 1919–1921 (the London bus engineer; GM’s motive per Wescott 1936; pp. 54, 56). The 1907–08 USGS tests stressed cleanliness. TEL was defended as “a certain means of saving petrol” (p. 53) and in 1973 as saving “one million barrels of oil a day” (p. 60). - The authors claim alternatives are “a useful defence against high oil prices” (p. 69). - The same macro-concern can be framed to support either side, depending on who controls the numbers.

15. Complexity, persistence and irreversibility. - Multiple lead sources complicate attribution (p. 47; Box 3.10 notes that solder and paint were removed at the same time). - Harms are largely irreversible (p. 47). Stocks persist in soils “for centuries” (p. 66). - The court allowed cumulative-source reasoning (p. 60). - The receding “safe” level (60 to 10 μg/dl, with no threshold apparent; p. 62) shows how the hazard’s apparent scale depended on the baseline available. Only as exposure fell could low-dose effects be measured (p. 61).

16. Innovation effects. - The industry frame treated TEL as the innovation and critics as anti-innovation (pp. 52–53). The authors contest this: other innovations (alcohol, synthetic fuels) were excluded from deliberation (p. 52; Panel 3.1). - According to Needleman’s recollection of an after-dinner conversation, a major producer (DuPont) declined R&D on a safer additive because of projected market decline (Box 3.9, p. 61). - Regulation later gave competitive advantage to carmakers with catalyst experience (p. 65). - The claim that a phase-out would have induced better innovation is a counterfactual assertion (p. 69).

Transferable insights (technology-neutral)#

1. Confident safety assurances are often given before any relevant data exist, and “no evidence of harm” is presented as “evidence of no harm”. - Evidence: Midgley 1922, “no actual experimental data has been taken” (p. 50); Standard Oil’s “no perils” (p. 51); lesson 3 (p. 70). - Strength: strong. Documented in contemporaneous letters. The authors note the same pattern in other Late lessons cases (p. 50).

2. When the sponsor controls research (funding, terminology, pre-publication review), the evidence base is shaped and narrowed for decades. Independent public funding is what eventually breaks the monopoly. - Evidence: Bureau of Mines terms (p. 50); 40 years of industry-only studies (p. 56); Kehoe’s lab (p. 56); the 1970 federal funding, which Needleman used for his low-dose studies (p. 60); society lessons 1–2 (p. 70). Note that Patterson, the first outside challenger, held PHS and petroleum-industry (API) contracts before 1965 (p. 58), so the “monopoly” was not absolute. - Strength: strong for the arrangements. Moderate for the causal claim that they delayed knowledge, which is inferred from sequence, though plausible.

3. Conditional approvals decay into unconditional ones. Safeguards and follow-up studies attached to an approval are not implemented unless mandated and funded, and voluntary industry compliance removes the pressure to create them. - Evidence: the committee’s caveats and its call for long-term study (p. 53); the PHS never funded it (p. 56); the 3 cc/gal voluntary compliance (p. 56); Table 3.1 “regulations, which do not take place” (p. 71). - Strength: strong. Clearly documented sequence, resting on Markowitz and Rosner (2002). Formally, only regulation was a proviso of the committee’s finding; the long-term study was strongly urged (pp. 53, 56).

4. Where the burden and standard of proof sit largely decides the outcome, and it is set by framing, rhetoric and institutional design rather than by the evidence alone. - Evidence: the 1925 afternoon debate and Howard’s reversal (pp. 52–53); the “precautionary statute” ruling (p. 60); the MRC question’s “prove … obvious harm” bar (p. 68). - Strength: strong that the burden and standard of proof were contested and moved in three independent episodes (US 1925, US courts 1976, UK 1980s). Moderate for the claim that this “largely decides the outcome”: the same chapter and panels show that economics and technological co-drivers also shaped outcomes (insight 13).

5. The endpoint chosen defines how big the hazard looks. Counting deaths or overt illness misses functional, subclinical and population-level harm. Assuming a threshold between “effects” and “adverse effects” delays prevention. - Evidence: Emerson (p. 52); Muskie and Patterson (p. 59); p. 60; lessons 2 and 9 (pp. 69–70); IQ effects at low levels (p. 61; Box 3.11, p. 62). - Strength: strong for lead. The generalisation to “conventional toxicology” (p. 70) is the authors’ broader claim and is contested in other domains.

6. Ubiquitous exposure hides harm. When comparison groups share the exposure, risk is underestimated, and widespread contamination gets relabelled as “normal” or “natural”. - Evidence: Kehoe’s Dayton “controls” and Mexican farmers (pp. 56–57); Patterson’s “typical” versus “natural” (p. 58); Bellow’s “zones of incomprehension” (p. 59); lesson 5 (p. 70). - Strength: strong. The methodological point is sound and was vindicated here.

7. The “safe” level recedes as measurement improves and background exposure falls. Early thresholds reflect the limits of available comparisons, not the absence of harm. - Evidence: CDC levels from 60 to 10 μg/dl (p. 62); comparisons with children at about 1 μg/dl (p. 61); Patterson’s contamination-free methods (p. 57). - Strength: strong for lead. Well documented regulatory history.

8. Concrete, near-term, concentrated benefits beat diffuse, long-term, uncertain harms in deliberation, and this can override the judgement even of experts who privately share the concern. - Evidence: p. 52 asymmetry; Hayhurst’s private letter versus his public stance (p. 53); society lesson 3 (p. 70). - Strength: moderate. Well illustrated by one documented individual and the meeting’s outcome. The general mechanism is inferred.

9. Undisclosed financial ties and dual roles (consultant-adviser, regulator-promoter) turn “independent” assurance into advocacy. - Evidence: Hayhurst (p. 52); Kehoe’s corporate roles (p. 56); Cumming’s promotion and chain of command to Mellon (p. 67). - Strength: moderate. Hayhurst is documented. The Cumming–Mellon link is hedged by the panel author (“may have had something to do”).

10. Framing a technology as essential to progress or national competitiveness recasts critics as reactionaries and narrows the options considered.

    • Evidence: industry’s four points (p. 52); Howard’s “gift of God” (p. 53); Hayhurst’s “survive among the nations” (p. 53); “No ‘innovation’ other than TEL was discussed” (p. 52).
    • Strength: moderate. Clear in the 1925 record. Its causal weight is the authors’ interpretation.

11. Once investment is committed, alternatives are denied and then forgotten. Proponents may disavow options they previously championed, and institutional memory of alternatives decays.

    • Evidence: Midgley 1921 versus Kettering/Midgley 1925 (p. 54); Winslow’s recommendation omitted (p. 53); the unpublished 1933 tests (p. 55); “forgotten by the 1960s” (p. 55).
    • Strength: moderate. A single panel author (Kovarik) with documentary quotations. Midgley’s USD 2/gal estimate (for a national supply of alcohol from farm wastes) and USGS’s “high cost” mean that denial of cost-competitive alternatives is less clearly shown than denial of technically feasible ones.

12. Market dominance entrenches a hazardous incumbent through distribution control, and even small cost frictions for alternatives in the supply chain may matter.

    • Evidence: Ethyl’s dealer restrictions and the antitrust findings (p. 55); the FTC order suppressing criticism (p. 55); Agrol’s one-penny handling cost (p. 55); the Standard–Farben market-division pact (p. 55).
    • Strength: moderate for market control, where the DOJ memo and court outcomes are documented. Suggestive for the supply-chain friction point: the Agrol account (Hale 1934; Kovarik) juxtaposes the extra penny and the plant’s closure without showing that one caused the other.

13. Final elimination often comes through an independent, coincidental driver (a technology compatibility issue or a separate environmental concern) rather than through health evidence alone. Such action is fragile if the driver disappears.

    • Evidence: GM and catalysts (p. 60); EPA’s push for a separate health standard (p. 60); forest die-back and “pure chance” (p. 64); the UK 1987 tax change “rather than in response to evidence of lead’s neurotoxicity” (p. 68); catalysts and price as uptake drivers (p. 65).
    • Counter-evidence: the first phase-downs (UK 1971, Germany 1972/1976) came “in response to scientific advice” (p. 63), and the US 1973 rule was health-framed and legally upheld on precautionary grounds (p. 60).
    • Strength: moderate to strong. Consistent across US, German/EU and UK accounts by different authors for the move to unleaded. It does not hold for the earlier health-based reductions.

14. Institutions can manage pressure without acting. Examples: another committee, targets set at levels already met, question framings that guarantee “unproven”, and monitoring that stops once policy changes (so outcomes cannot be evaluated).

    • Evidence: Millstone’s UK account (pp. 67–68); the delayed UK phase-down (p. 63).
    • Strength: moderate. Well argued by one analyst. It is his interpretation of motives, though the survey-based action level is a concrete example.

15. Cost projections by affected industries can greatly exceed the costs realised, and unrebutted cost objections can delay regulation.

    • Evidence: “one million barrels a day” against the EPA’s 30 000 (p. 60); German macro costs “insignificant” and claimed costs “significantly biased” (p. 65). Separately, the EP committee postponed the 0.15 g/l stage because industry’s investment and fuel-consumption objections “could not be refuted” (p. 63). That shows cost claims carrying weight, not that they were inflated.
    • Strength: moderate. One ex-post economic study (Hagner 2000) plus one advertising claim. The chapter does not say what the advertisement’s figure measured, so the “33-fold” gap may not be like-for-like. Generalising to “systematically” is not supported by so few data points.

16. A statute that explicitly authorises anticipatory action and cumulative-exposure reasoning lets regulators act before harm is proven, and courts may uphold that.

    • Evidence: Ethyl Corp. v EPA (p. 60).
    • Strength: moderate. A single, important legal precedent, quoted directly.

17. Scientists who raise early warnings commonly face loss of funding, threats to employment or attacks on character, and vindication arrives decades later.

    • Evidence: Baker (p. 48); Henderson (p. 51); Patterson (p. 58); Needleman (p. 61); society lesson 5 (pp. 70–71).
    • Strength: moderate. The pattern recurs, but several accounts rest on the protagonists’ own testimony (Patterson’s letter to Needleman; Needleman on himself), and retaliation is sometimes inferred from timing. The authors’ lesson also lists Byers (p. 70), whose harassment the chapter does not document.

18. Harm from a dispersed persistent agent outlives its use by generations, and its costs shift onto the public purse and the most vulnerable.

    • Evidence: legacy soil and dust (pp. 47, 62, 66, 69); US remediation and litigation (p. 49); irreversibility (p. 47); e-waste in low-income countries (p. 69).
    • Strength: strong. Well documented for lead.

19. Small average individual effects across a whole population can produce large social harm at the tails of a distribution. Dismissing them as “small” is an analytical error.

    • Evidence: the 5-point IQ shift and the doubling and halving of the tails (p. 61); USD 100–300 billion per cohort (p. 62).
    • Strength: strong. The statistical logic holds under a normal distribution. The economic estimate comes from one cited study.

20. Hazards that can be controlled in a bounded setting cannot be controlled once dispersed through society. Evidence from robust, highly exposed groups does not bound risk to sensitive groups under chronic multi-route exposure.

    • Evidence: Hamilton, “how can you control the whole country?” (p. 53); science lesson 1 (p. 69).
    • Strength: moderate. A strong argument, illustrated rather than tested here.

21. A decision taken in one leading jurisdiction becomes a global default through expert assurance, licensing and trade. Phase-out in stricter markets can push sales to weaker jurisdictions.

    • Evidence: the UK’s 1930 reliance on US experts (p. 67); German licensing (p. 57); the Octel bribery convictions (p. 68); exposure persisting in poorer countries (pp. 47, 69).
    • Strength: moderate for the export of the 1925 decision (p. 67). Suggestive for the second sentence: it is an inference from one bribery case, reported by a newspaper (Leigh et al. 2010) whose details should be checked. The chapter itself does not make the displacement generalisation.

22. A determined frontrunner jurisdiction can pull a multi-state system to higher standards, and firms already adapted to stricter rules gain competitive advantage.

    • Evidence: Germany in the EC (p. 63); German carmakers with US catalyst experience (p. 65); Wedgwood’s “level playing field” (p. 48).
    • Strength: moderate. Documented by Haigh and von Storch for one case.

23. Timing matters: well-organised campaigns and political windows can convert standing expert advice into rapid policy change.

    • Evidence: CLEAR plus the election, with acceptance in half an hour (p. 63).
    • Strength: suggestive to moderate. One episode, although Haigh states it “with some certainty”.

24. Suppressing hazard vocabulary is a deliberate tactic that shapes public perception.

    • Evidence: “ethyl” not “lead”, including in internal correspondence (p. 50); “undue” rather than “poisoning” (p. 60); MRC’s “prudent” (p. 68).
    • Strength: moderate. Documented instances. Intent is sometimes inferred.

25. Leaders of a dominant firm may make product decisions on projected market economics, while public debate is conducted on scientific grounds as a proxy.

    • Evidence: Box 3.9, DuPont’s declined USD 100m R&D and the “shadow play” (p. 61).
    • Strength: suggestive. A single, informal, recalled conversation reported by an interested party.

Limitations, contestation and bias check#

Advocacy rather than analysis - The chapter is openly a precautionary morality tale. The section title “‘Progress’ or precaution?” (p. 52) sets the two against each other. - Industry actors are characterised with loaded language (“plumbing new depths”, p. 57; “unconscious irony”, p. 60; “Apparently, poisoning a technology was more important than poisoning people”, p. 60). - Motives are frequently inferred from timing, for example the termination of Henderson’s contracts (p. 51) and of Patterson’s (p. 58). - The lessons (pp. 69–71) generalise from this single, clearly vindicated case to “conventional toxicology” and “public health” at large.

Author-as-protagonist - Needleman co-wrote a history in which he is a central figure, and his annex bio credits his advocacy with the bans (p. 695). - His work is called “seminal” and paradigm-shifting (p. 60). Industry’s response is “character assassination” (p. 61), with no description of what that involved or how it was resolved. - Box 3.9 is his own after-dinner anecdote. - The only source cited for Patterson’s retaliation episode is his 1992 letter to Needleman (p. 58; ref. p. 74), and only for the “buy me out” quote. The contract losses and trustees’ visit are uncited. - This does not make the account wrong, since much is corroborated by the later science and by independent histories. But it is not a disinterested history.

Heavily secondary sourcing - The 1920s–60s narrative follows Markowitz and Rosner (2002), Needleman’s essays (1997, 1998, 2000), Kovarik (2003, 2005) and Davis (2007). - Archival letters are cited, but the chapter gives no sign they were re-examined. - No industry-side history, Kehoe defence or contemporary regulator’s account is represented. - The front matter claims case chapters were peer reviewed “from a broad range of perspectives” (PDF p. 7), but no dissenting panel appears. - External note, to verify: Markowitz and Rosner served as expert witnesses in lead litigation, which is relevant to how one weighs a single-source history.

Hindsight bias - The 1925 critics are judged “prescient” (p. 52) because they turned out right. - The chapter does not examine what a Henderson-style standard (“absolute proof … that no danger”, p. 51) would imply if applied generally. Nor does it ask how decision-makers in 1925 could have told this warning apart from warnings that proved unfounded. - Hayhurst’s “thousand and one other poisons” worry (p. 53) is dismissed as politically motivated rather than engaged with as a real policy problem. - The chapter’s own evidence shows that the population-scale low-dose harm was not demonstrable with 1925 methods (p. 61).

The alternatives claim is overstated in places - The summary says the alcohol additive was “equally effective” and “cleaner” (p. 46). Panel 3.1’s evidence supports technical feasibility and anti-knock equivalence (the 1933 tests, p. 55). - But Midgley’s own estimate put alcohol from farm wastes, at national scale, at USD 2/gal (p. 54), and USGS noted its “high cost” (p. 54). Box 3.6 has petrol “slightly ahead” on mileage (p. 56). The chapter gives no systematic cost, scale or supply comparison. - The rejection motive is “likely … less profitable” in the text (p. 50) but hardens to “rejected as less profitable” in Table 3.1 (p. 71). - The only acknowledged substitute hazard is benzene (p. 56). - External to the chapter (to verify): alcohol fuel faced Prohibition-era constraints in 1920–33. Post-phase-out substitutes carried their own risks; Late lessons Vol. 1 (EEA 2001) had a chapter on MTBE as a lead substitute that polluted groundwater. The chapter does not address regrettable substitution.

Counterfactuals asserted - “A 10-year phase out … at any time since 1925 would have encouraged innovators” (p. 69) and “benefits could … have been attained by alternative uses of the economic capital” (p. 69) are plausible but not argued with evidence.

The chapter’s own evidence complicates its science-to-policy narrative - The chapter candidly shows that the final move to unleaded came via catalytic converters, forest die-back, elections and “pure chance” (pp. 60, 63–64, 68). - The US phase-down rule (1973) preceded Needleman 1979. - In the other direction, the earlier phase-downs (UK 1971, Germany 1972/1976) followed scientific advice (p. 63), and the US rule was health-framed and upheld as precautionary (p. 60). - Overall, this supports a “contingency and co-drivers” reading at least as much as a “precaution vindicated” one. The chapter’s concluding lessons do not draw this out; the panels do.

Internal inconsistencies and factual slips (to verify before relying on the specific numbers or dates) - Peak emissions: “about 200 000 tonnes … in both the US and Europe” (p. 46), against Europe’s peak of 119 000 t in 1975 (p. 65). For the US, 250 000 t (p. 56) against 200 000 t (p. 60). - Patterson’s multiple: “600 times” (p. 57), “100 times” (Table 3.1, p. 71), “hundreds of times” (pp. 46, 70). - Toxic threshold of 60 μg/dl: “In the 1970s” (p. 61) against “In the 1960s” (Box 3.11, p. 62). - US phase-out given as 1995 (p. 62). The on-road ban is usually dated 1 January 1996 (external, to verify). - Aarhus: the text cites the “Aarhus Protocol”, but the reference (p. 75) is the Aarhus Convention on access to information. The EU’s own leaded-petrol ban came via Directive 98/70/EC, which the chapter does not mention (external, to verify). - Franklin: “In 1818, Benjamin Franklin, while ambassador to France” (p. 49). Franklin died in 1790, so 1818 is the publication date. - Wedgwood: “Josiah Wedgewood” and the 1833 Factories Act (p. 48). The founder died in 1795. A later family member may be meant, or the anecdote is garbled. - Conference date: 24 May 1925 (p. 52). Standard histories give 20 May 1925 (external, to verify). - Ban: “New York State then banned the sale” (p. 51). Standard accounts describe bans by New York City and other localities (external, to verify). - Ethyl’s founders: “the three companies created the Ethyl Corporation” (p. 50). Ethyl is usually described as a GM–Standard Oil of New Jersey joint venture, with DuPont as manufacturer and GM shareholder (external, to verify). - Kehoe’s affiliations: “corporate officer at GM” (p. 56). He is usually described as Ethyl’s medical director (external, to verify). - Octel: “in the summer of 2010 two former senior executives of Octel were convicted” (p. 68). External recollection, to verify: the company (Innospec) pleaded guilty in 2010, and individual executives were convicted or pleaded guilty later (c. 2012–2014). - Needleman 1979 subjects: “milk teeth from more than 2 000 infants” (p. 61). The subjects were schoolchildren, and the IQ difference reported in the 1979 paper is usually given as about 4 points between high and low dentine lead groups (external, to verify). - Table 3.1 misdates the temporary suspension to 1923–24 and has “for next 50–60 years” against “1930–1960s”. It gives “low-lead” where Panel 3.3 says unleaded. - Names and typos: “Huge Cummings” (Hugh Cumming), “Midgely”/”Midgley”, “Craig Patterson” (Clair), “Yandall”, “leaded patrol”, “European Food Standards Agency” (EFSA is the European Food Safety Authority), “WOBLIP”/”WOPLIP”, “Kitmann”/”Kitman”, “Contentsies” (p. 56). - Box 3.12 duplicates two sentences from Box 3.11. - Overstatement: “For the next 40 years all studies of TEL were conducted and funded by the Ethyl Corporation and GM” (p. 56). The chapter itself says Patterson had a “longstanding contract with the Public Health Service” and a “substantial contract with the American Petroleum Institute” before 1965 (p. 58). His work was geochemical rather than a TEL health study, but it is what challenged the TEL paradigm. The Byers and Lord 1943 study (p. 61) was on childhood lead poisoning, not TEL. - Muskie: the chapter calls him “Senator Edward Muskie” (p. 58). It was Edmund Muskie. - John Snow: the cholera work is dated “1864” (p. 58). Snow’s cholera studies date from 1849–1854 (external, to verify). - Devon colic: described as “early-18th century” (p. 48), but Baker’s paper is dated 1768. - Mansfield Clark: the chapter calls him “a professor of chemistry” (p. 50). The reference routes his memo “through Acting Director, Hygenic Laboratory” (p. 73), and external sources describe W. Mansfield Clark as a PHS Hygienic Laboratory chemist in 1922 (external, to verify). - Alternatives at the conference: “No ‘innovation’ other than TEL was discussed” (p. 52), against Panel 3.1’s “except in a few statements by public health scientists and hints in the media” (p. 55). - MRC 1988 is characterised two ways by Millstone: a hypothesis “consistent with” a small effect (p. 67) versus “accepted” a small negative effect (p. 68). - Needleman’s decade: “Needleman in the 1970s” (society lesson 1, p. 70) against “Needleman in the 1980s” (society lesson 5, p. 70). - Table 3.1: “1920 Leaded paint is banned in Australia and later in Europe”, against Box 3.4’s 1909–1930s bans in Europe and Australia (p. 49). “1995–2000 Virtually all western Europe only uses lead-free petrol”, against Panel 3.3’s under 30% unleaded share in France, Greece and Portugal in 1995 (p. 65). - Reference typos: EEA 2001 is given as “1986–2000” (should be 1896–2000; p. 72); “Chishold, 1976” in Panel 3.4 for Chisholm (p. 67).

Generalisations to contested contemporary issues - Footnote 3 (p. 50) uses the “no evidence of harm” lesson to assert risks to children from mobile phones. - The “shooting the messenger” passage (p. 58) lists climate change, GMOs and electromagnetic fields together as fields where scientists who publish warnings are attacked. Strictly, the point concerns how messengers are treated, not whether each warning is correct. But the grouping implies an equivalence the chapter does not examine, and the evidence bases in these areas differ. - This is the clearest example of the pro-precaution framing driving conclusions beyond the case evidence.

Treatment of opponents - Kehoe is presented almost wholly as a captured authority. His 1965 review letter (p. 58) is damning in tone. The chapter records one concession, that he “eventually came to see the merit in his critics’ assertions” about his control group (p. 56). It acknowledges that the threshold assumption is “still a dominant assumption in conventional toxicology” (p. 70). But it does not acknowledge any contribution from his metabolic studies or engage with why the threshold view was, and is, mainstream. - The panels by Haigh and von Storch are more measured. Haigh notes that initial action in both countries came “in response to scientific advice, without much public pressure” (p. 63) and that EC experts saw “no immediate danger” (p. 63). This is a useful corrective to a pure “capture” narrative. - Millstone is critical of UK scientists’ caution but hedges on the Octel causal link (“difficult to establish”, p. 68).

Case selection - Lead in petrol is one of the clearest cases in which early warnings were correct, the harm was large and the alternatives were feasible. It is well suited to illustrating precautionary lessons and poorly suited to testing them. False-positive warnings about fuel additives are not considered.

Fair in the other direction - The core empirical story is robust: - lead is neurotoxic without a clear threshold (Box 3.1, Box 3.11); - removing lead from petrol was followed by large falls in blood lead, tracked by air lead (EPA’s causal finding, p. 61; Box 3.10); - industry cost claims were overstated (pp. 60, 65); - the conditions attached to the 1926 approval were not implemented (p. 56). - The chapter openly acknowledges several counterweights: the benefits of leaded petrol (p. 69); that solder and paint were removed at the same time (Box 3.10); that attribution across multiple sources is difficult (p. 47); that other public health scientists supported industry (p. 52); and the committee’s own caveats (p. 53).

Notable quotes#

1. “conditions will grow worse so gradually, and the development of lead poisoning will come on so insidiously … that leaded petrol will be in nearly universal use … before the public and the government awakens to the situation” (Henderson 1925; p. 47, p. 52)
2. “very serious consideration … although no actual experimental data has been taken” (Midgley 1922; p. 50)
3. “‘No evidence of harm’ is thereby mischaracterised as ‘evidence of no harm’.” (authors; p. 50)
4. “You may control conditions within a factory … but how can you control the whole country?” (Alice Hamilton 1925; p. 53)
5. “Because some animals die and some do not die in some experiments, shall we give this thing up entirely?” (Frank Howard, Ethyl, 1925; p. 53)
6. “I am afraid human progress cannot go on under such restrictions … if we are to survive among the nations.” (Hayhurst 1925, private letter; p. 53)
7. “Classical poisoning is just one extreme of a whole continuum of responses of an organism” (Patterson 1966; p. 59)
8. “the regulatory action under this precautionary statute [the Clean Air Act] should precede, and hopefully prevent, the perceived harm.” (Ethyl Corp. v EPA 1976; p. 60)
9. “It is pure chance that they came together at the same time” (Haigh, on health and forest die-back; p. 64)
10. “It is often more convenient for a hazardous industry to debate the science than to discuss options for reducing hazards.” (authors; p. 70)

Open questions#

1. Could a 1925 decision-maker have told this warning apart from ones that proved unfounded? The chapter implies the signals were plain: known toxicant, cumulative, worker deaths, available alternatives. It does not state criteria. Chapter 26 (science for precautionary decision-making) may supply them.
2. How robust is the alternatives case on cost and scale? What would a 1920s ethanol or blend pathway have cost, given Prohibition-era alcohol controls and supply? What hazards did post-phase-out substitutes carry (cross-check the MTBE chapter in EEA 2001)?
3. How much of the phase-out was health-driven versus catalyst-driven, in each jurisdiction and over time? Would lead have returned with a lead-tolerant catalyst, as the EPA staff and Haigh feared (pp. 60, 64)? Were health standards eventually made independent of catalyst compatibility?
4. What exactly was the “character assassination” of Needleman (p. 61), and how was it resolved? The chapter’s silence matters for assessing the harassment lesson fairly. Were the criticisms of his 1979 methods purely strategic, or partly methodological?
5. Primary archival verification: the Bureau of Mines contract terms, Winslow’s omitted recommendation, the unpublished 1933 military tests, and the Commerce report “published but not released”. These rest on Kovarik and on Markowitz and Rosner.
6. Does the low-dose, supralinear dose–response (Lanphear 2005; p. 62) still hold in later pooled analyses? Where did CDC and EFSA reference values go after 2012?
7. Did the lead–antisocial behaviour link (p. 61) strengthen or weaken with later, better-controlled studies?
8. Octel/Innospec: what were the precise dates and outcomes of the bribery cases? When did the last countries stop selling leaded petrol, and did any leaded fuel uses remain outside road transport?
9. Did UK child blood lead surveillance resume after being stopped (p. 68), and what did it show?
10. Generalisability: the authors extend the lessons to mobile phones, GMOs and EMF (pp. 50, 58). Which transferable insights hold only for persistent, bioaccumulative, dose-dependent toxicants, and which are about institutions and knowledge production and so travel more widely?

Audit log#

Independent audit against the full text extract (PDF pp. 48–77), the reference list, the author annex (PDF pp. 691–701) and front matter (PDF p. 7). Figures 3.1 and 3.2 were re-checked visually. Changes to these notes:

1. Section 3.3 heading corrected to the source title. Added the chapter’s statement that household paint was the main public lead source until 1925 (p. 50).
2. Mansfield Clark: now uses the chapter’s own description (“a professor of chemistry”), with the reference-list date (11 October 1922) and Hygienic Laboratory routing (p. 73).
3. Henderson: the chapter labels harassment outright, on timing evidence only. Was “implies retaliation”.
4. Added that the Surgeon General called the 1925 conference in response to “requests for action from public health experts” (p. 52).
5. Added Edsall’s rejection of industry’s “nobody has shown any symptoms” claim (p. 52).
6. Hayhurst/Thompson quote made verbatim; added Hayhurst’s “Personally, I can quite agree” (p. 53).
7. Committee finding: the only formal proviso was “proper regulations”. Long-term study was urged, not a stated condition (Table 3.1 differs). Applied in 3.4.2, mechanism 3, insight 3 and the digest.
8. Panel 3.1: noted that alternatives surfaced at the conference only in “a few statements” (p. 55), against the chapter’s “No ‘innovation’ … was discussed” (p. 52).
9. Agrol: softened from “was enough to kill the alternative” to an implied, unproven link, in section notes, mechanism 8 and insight 12. Insight 12’s friction claim rated suggestive.
10. Added Kehoe’s recorded concession on his control group (p. 56), in section notes and the treatment-of-opponents paragraph. The latter also now notes the chapter’s acknowledgement that threshold thinking is “still a dominant assumption” (p. 70).
11. Added the Ethyl Gemeinschaft / I.G. Farben 1930s context (p. 57).
12. Shooting-the-messenger comment reworded. The point concerns treatment of messengers; the criticism is that the grouping implies equivalence. Noted the “1864” Snow date.
13. Patterson retaliation: the only cited source (1992 letter to Needleman) supports only the “buy me out” quote. Contract losses and the trustees’ visit are uncited, and the link to Ethyl is conveyed by sequence.
14. Noted the chapter’s misnaming “Edward Muskie” (p. 58).
15. Added that Needleman used the new public funds (p. 60), in section notes and insight 2.
16. Box 3.10: corrected the claim that the box “notes” confounding. It states the co-occurrence only. Noted that the IQ gain is tied to blood lead decline, not petrol specifically.
17. Panel 3.2: added the EP committee’s approval of the 0.4 g/l first stage, and that 0.15 g/l was the Directive’s lowest permissible level.
18. Figure 3.1: the stepped line is unlabelled (values match the UK); full caption given.
19. Panel 3.3: added “preliminary” to the press analysis; μg/L to μg/dl conversion for Figure 3.2; model-based nature of the “reduced health hazards” claim; undefined “critical level”; uncited “significantly biased” cost claim.
20. Panel 3.4: flagged Millstone’s two characterisations of MRC 1988 (hypothesis on p. 67, “accepted” on p. 68). Added the “marginally modified” note.
21. Table 3.1: added the 1920-versus-1909 paint ban and 1995–2000 all-unleaded discrepancies.
22. Timeline: flagged 1786 as an external date; corrected “not cost” (USGS noted “high cost”); used the chapter’s name for Ethyl; noted the undated first court ruling.
23. “What was known”: Clark’s street-dust point reframed as a prediction, with the committee’s still-open garage dust question (p. 53).
24. The US 1973 phase-down is no longer described as “driven by” catalyst compatibility. It was health-framed, with EPA staff seeking a catalyst-independent standard (p. 60).
25. Science lesson 4 (“balance”) flagged as absent from the chapter body and moved from “derived from case evidence” to “thinner support”. The attribution to Kehoe in mechanism 5 is marked as this note’s inference.
26. Science lesson 7: refined the “no examples” comment (Bureau of Mines study; Howard’s remark).
27. Society lesson 3: noted “gift from God” is attributed to “the car industry” here but to Ethyl’s president on p. 53.
28. Society lesson 5: Byers’s harassment is undocumented in the chapter; Needleman is dated to both the 1970s and the 1980s. Carried into insight 17 and the digest.
29. Lesson 8: support identified as Kehoe episodes plus the Box 3.9 recollection; the lesson’s own hedging noted.
30. Mechanism 2: “Timed reassurance” softened; no deliberate timing is claimed by the chapter.
31. Mechanism 7: noted that the chapter also lists diffuse benefits (efficiency, jobs; p. 69).
32. Alternatives: Midgley’s USD 2/gal clarified as the cost of a national cellulose-alcohol supply, not like-for-like with the 1933 blend tests. Applied in mechanism 9, insight 11, the bias check and the digest.
33. Mechanism 10 and insight 13 limited to the final move to unleaded. Added counter-evidence (UK 1971, Germany 1972/76 on scientific advice; the health-framed US rule). Insight 13 strength lowered from strong to moderate to strong.
34. Mechanism 12: “unanimity let a frontrunner shape the floor” corrected. Directive 78/611 allowed national 0.15 g/l limits under a 0.4 g/l cap. Octel convictions were of two former executives, not the firm. The displacement generalisation is marked as inference.
35. Mechanism 13: the press did not “force” the conference. The chapter credits requests from public health experts (p. 52).
36. Mechanism 14: the alcohol-as-oil-hedge dating corrected from 1907 to 1919–21 (the USGS tests were about cleanliness).
37. Mechanism 16: Box 3.9 DuPont R&D claim marked as Needleman’s recollection.
38. Insight 4: strength split (strong that the burden was contested; moderate as the decisive factor).
39. Insight 15: removed the EP committee as evidence of inflated costs, since it shows unrefuted objections carrying weight. Noted the ad and EPA figures may not be like-for-like; dropped “systematically”.
40. Insight 21: the second sentence is rated suggestive and marked as inference.
41. Bias check: Patterson sourcing made precise; alternatives paragraph adds USGS “high cost” and Box 3.6 mileage; contingency paragraph balanced with the early health-based reductions.
42. Slips list: replaced the self-cancelling Byers “overstatement” with the Patterson PHS/API contracts. Added Muskie, Snow 1864, Devon “early-18th century”, Mansfield Clark description, the alternatives-discussed tension, MRC 1988, Needleman’s decade, two Table 3.1 conflicts, and reference typos (EEA 2001 “1986–2000”, “Chishold”).
43. Mechanism 11: the UK action level is described in Millstone’s terms rather than as intentionally “chosen” to require no action.

Verified as accurate and left unchanged: author and panel bios and page numbers (annex pp. 689–699), the front-matter peer review statement (PDF p. 7), all numerical data checked (emissions, blood lead, CDC levels, IQ and cost figures, dates of directives), the Figure 3.2 readings, the notable quotes, and existing flags on the Aarhus citation, Franklin, Wedgwood, the 600× versus 100× discrepancy, and the Europe 200 000 t versus 119 000 t conflict.