# Lateral library — the "numerology heirs" papers (held locally, 2026-06-11)

*Primary/authoritative PDFs for the five lateral components surfaced in
`_workflow_results/LATERAL_GAPS_2026-06-11.md` (the 10⁻⁶ carcinogen de minimis; Bruce Ames; the RfD
uncertainty-factor tower; GRAS self-certification; TLVs). Each was **title-verified** against its intended
citation (one bad fetch was caught and replaced — Europe PMC's PMC433250 is a sulfenyl-iodide paper, not
Ames 1973) **and content-verified** (the load-bearing quote was grepped inside the file). Every item here was
obtained via **open-access / government routes only** (PMC/Europe PMC, PNAS open archive, govinfo, GAO via
Wayback, EPA, NRDC) — **no shadow libraries**, per CLAUDE.md. Paywalled targets are in `_WISHLIST.md`.*

**Grades:** [CONFIRMED-primary] = primary read locally · [secondary] = authoritative non-primary.
Each `.pdf` has a `-layout` `.txt` sidecar.

## A. The 10⁻⁶ "one in a million" carcinogen de minimis (the headline gap)
- **✓ Mantel & Bryan (1961)** — `mantel-bryan1961_jnci_safety-testing-carcinogenic-agents.pdf` (+`.txt`).
  Nathan Mantel & W. Ray Bryan (NCI Biometry Branch), "'Safety' Testing of Carcinogenic Agents," *J. Natl.
  Cancer Inst.* 27(2):455–470. DOI 10.1093/jnci/27.2.455. **[CONFIRMED-primary] — operator Taildrop
  2026-06-11.** The **original "virtually safe dose"** at a lifetime risk of **1/100 million (10⁻⁸)** — the root
  Hutt later moved up to 10⁻⁶ (Rodricks). Verified inside: *"How safe is safe? Absolute safety can never be
  unquestionably demonstrated…"*; *"a risk of 1/100 million is so low as to constitute 'virtual safety.' **Other
  arbitrary definitions of 'virtual safety' may be employed as conditions require.**"* The Summary calls the
  definition, the assurance level, and the extrapolation slope all "**arbitrary**." Excerpt:
  `sources/MantelBryan_1961_virtually-safe-dose_excerpt.md`.
- **✓ Rodricks (2019)** — `rodricks2019_risk-assessment-came-to-washington.pdf` (+`.txt`). Joseph V. Rodricks,
  "When Risk Assessment Came to Washington: A Look Back," *Dose-Response* 17(1):1–15. DOI
  10.1177/1559325818824934 (PMC6366000). **[CONFIRMED-primary — firsthand participant memoir].** Verified
  inside: the **1962 DES Proviso**; **Mantel's risk target "1 in 100 million," labeled "virtually safe"**; and —
  the prize — *"after extensive discussions with me and other scientists, Hutt proposed that 'safe doses' for
  carcinogens such as DES could be defined as those associated with lifetime risk levels of less than 1 in 1
  million, when these risks were estimated using a linear, no-threshold model."* This is the 10⁻⁶ origin from a
  man who was in the room. Excerpt: `sources/Rodricks2019-Crump2018_10-6-origin_excerpt.md`.
- **✓ Crump (2018)** — `crump2018_doseresponse_VSD-history.pdf` (+`.txt`). Kenny Crump, "Cancer Risk Assessment
  and the Biostatistical Revolution of the 1970s—A Reflection," *Dose-Response* 16(4):1–8. DOI
  10.1177/1559325818806402 (PMC6295697). **[CONFIRMED-primary].** Verified inside: the **single-cell / no-
  threshold** realization that broke the NOAEL-safety-factor model; the **Mantel-Bryan (1961) log-probit
  procedure** and its "virtually safe dose" (Fig. 1), "apparently … never used by a regulatory" agency as such.

- **✓ Rulis (1992)** — `rulis1992_acs484_threshold-of-regulation.pdf` (+`.txt`). Alan M. Rulis (FDA),
  "Threshold of Regulation: Options for Handling Minimal Risk Situations," ch. 14 in *Food Safety Assessment*,
  ACS Symp. Ser. 484, pp. 132–139. DOI 10.1021/bk-1992-0484.ch014. **[CONFIRMED-primary] — operator Taildrop
  2026-06-11.** The **477-carcinogen potency-distribution** derivation ("Ref. 1" of the 1995 ToR rule) — the
  calculation that makes 0.5 ppb a sub-10⁻⁶ level. Same Rulis who credited "Frawley 1967" (register D5).
  Excerpt: `sources/Rulis1992_477-carcinogen-threshold_excerpt.md`.
- **✓ Kelly (1991, updated)** — `kelly1991_myth-of-10-6-acceptable-risk.pdf` (+`.txt`). Kathryn E. Kelly,
  "The Myth of 10⁻⁶ as a Definition of Acceptable Risk" (A&WMA 84th Ann. Mtg., Vancouver, 1991; held copy is the
  author's updated version). **[CONFIRMED-primary]** as read — the authoritative investigation: 10⁻⁶ "has **no
  scientific or regulatory basis**." Carries the officials' confessions ("sounded like such a nice phrase,"
  "doable," "You really shouldn't be asking these questions") and the Mantel **"we just pulled it out of a hat"**
  line (Kelly = the source; grade the anecdote **[secondary]**). Independently corroborates the
  cranberry→Mantel-Bryan→DES/SOM→10⁻⁶ chain. Excerpt: `sources/Kelly_1991_myth-of-10-6_excerpt.md`.
- **✓ Travis, Richter, Crouch, Wilson & Klema (1987)** — `travis-etal1987_cancer-risk-management-132-decisions.pdf`
  (+`.txt`). *Environ. Sci. Technol.* 21(5):415–420. DOI 10.1021/es00159a001. **[CONFIRMED-primary]** (via
  `doc_retrieve`/scidb; register H24). The **empirical test of 10⁻⁶**: across **132 federal decisions** the
  *de facto* acceptable risk tracked **nearer 10⁻⁴**, varying with exposed-population size (political, not
  toxicological); recites "de minimis non curat lex." Companion to Kelly — 10⁻⁶ isn't even what agencies apply.

## B. Bruce Ames — de minimis scaled to the whole diet
- **✓ Ames, Durston, Yamasaki & Lee (1973)** — `ames1973_pnas_carcinogens-are-mutagens.pdf` (+`.txt`).
  "Carcinogens are Mutagens: A Simple Test System Combining Liver Homogenates for Activation and Bacteria for
  Detection," *PNAS* 70(8):2281–2285. DOI 10.1073/pnas.70.8.2281 (PMC433718). **[CONFIRMED-primary].** The
  **Ames test** — the no-threshold (mutagen→carcinogen) screening tool. The irony engine of the section.
- **✓ Ames, Profet & Gold (1990)** — `ames1990_pnas_dietary-pesticides-99-99-natural.pdf` (+`.txt`). "Dietary
  pesticides (99.99% all natural)," *PNAS* 87(19):7777–7781. DOI 10.1073/pnas.87.19.7777 (PMC54831).
  **[CONFIRMED-primary].** Verified inside: *"99.99% (by weight) of the pesticides in the American diet are
  chemicals that plants produce to defend themselves"*; ~1.5 g/day natural vs synthetic residues; *"the
  comparative hazards of synthetic pesticide residues are insignificant."* The round number that universalizes
  the trifle. Excerpt: `sources/Ames_1973-1990_natures-pesticides_excerpt.md`.
- **✓ Ames & Gold (1990)** — `ames-gold1990_pnas_too-many-rodent-carcinogens.pdf` (+`.txt`). "Chemical
  carcinogenesis: Too many rodent carcinogens," *PNAS* 87(19):7772–7776. DOI 10.1073/pnas.87.19.7772
  (PMC54830). **[CONFIRMED-primary].** "Half are rodent carcinogens"; the MTD/mitogenesis attack on the rodent
  bioassay — "natural ≠ safe cuts both ways."

- **✓ Ames, Magaw & Gold (1987)** — `ames-magaw-gold1987_science_ranking-carcinogenic-hazards.pdf` (+`.txt`).
  "Ranking Possible Carcinogenic Hazards," *Science* 236(4799):271–280. DOI 10.1126/science.3563506.
  **[CONFIRMED-primary] — operator Taildrop 2026-06-11.** The **HERP** index ("Human Exposure dose/Rodent
  Potency dose," Table 1) — the peanut-butter-sandwich/wine ranking that makes synthetic residues look trivial.
  (Free mirror: `files.toxplanet.com/cpdb/pdfs/Science-Ranking.pdf`.) Folds into
  `sources/Ames_1973-1990_natures-pesticides_excerpt.md`.

## C. The RfD / uncertainty-factor tower (the factor of 100, metastasized)
- **✓ U.S. EPA (2002)** — `epa2002_RfD-RfC-process-review.pdf` (+`.txt`). Risk Assessment Forum, *A Review of
  the Reference Dose and Reference Concentration Processes*, EPA/630/P-02/002F (Dec 2002), 192 pp.
  **[CONFIRMED-primary].** Verified inside: the standard **10-fold default UFs** — interspecies, intraspecies,
  subchronic→chronic, LOAEL→NOAEL, database — and the RfD as an estimate carrying uncertainty "of an order of
  magnitude." The official method built from stacked tens. Excerpt:
  `sources/EPA2002_RfD-uncertainty-factor-tower_excerpt.md`.

## D. GRAS self-certification (the mechanism Frawley was actually working)
- **✓ U.S. GAO (2010)** — `gao-10-246_2010_GRAS-oversight.pdf` (+`.txt`). *Food Safety: FDA Should Strengthen
  Its Oversight of Food Ingredients Determined to Be Generally Recognized as Safe (GRAS)*, GAO-10-246 (Feb
  2010). **[CONFIRMED-primary].** Verified inside: companies may self-determine GRAS *"without FDA's approval or
  knowledge"* and *"are not required to inform FDA of them."*
- **✓ FDA (2016)** — `fr2016_GRAS-final-rule_81FR54960.pdf` (+`.txt`). *Substances Generally Recognized as
  Safe*, final rule, 81 FR 54960 (17 Aug 2016); Docket FDA-1997-N-0020. **[CONFIRMED-primary].** Confirms the
  GRAS-notification procedure is **voluntary**.
- **✓ NRDC (2014)** — `nrdc2014_generally-recognized-as-secret.pdf` (+`.txt`). Tom Neltner & Maricel Maffini,
  *Generally Recognized as Secret: Chemicals Added to Food in the United States*, NRDC R:14-03-A (Apr 2014).
  **[secondary — advocacy NGO; grade contested estimates as such].** Identifies **275 chemicals / 56 companies**
  on undisclosed GRAS; cites an estimate of **~1,000 such secret determinations**. GRAS excerpt:
  `sources/GRAS_self-determination_GAO-NRDC-FR2016_excerpt.md`.

- **✓ Neltner et al. (2013)** — `neltner2013_jama_gras-conflicts-of-interest.pdf` (+`.txt`; regulations.gov
  mirror also held: `neltner2013_jama_gras-coi_regulationsgov-mirror.pdf`). "Conflicts of Interest in Approvals
  of Additives to Food Determined to Be GRAS: Out of Balance," *JAMA Intern. Med.* 173(22):2032–2036. DOI
  10.1001/jamainternmed.2013.10559. **[CONFIRMED-primary] — operator Taildrop 2026-06-11.** Of **451** GRAS
  notifications (1997–2012): **22.4%** assessed by a manufacturer employee, **13.3%** by a manufacturer-selected
  consulting firm, **64.3%** by a manufacturer/firm-selected expert panel, **0%** by an independent third party.
  GAO+NRDC+FR2016 carry the structure; **JAMA carries the numbers.**

## E. TLVs (occupational mirror) — NOW HELD (via `doc_retrieve`, 2026-06-11)
- **✓ Castleman & Ziem (1988)** — `castleman-ziem1988_corporate-influence-on-TLVs.pdf` (+`.txt`). *Am. J. Ind.
  Med.* 13(5):531–559. DOI 10.1002/ajim.4700130503. **[CONFIRMED-primary].** "Unpublished corporate
  communications were important in developing TLVs for **104 substances; for 15 of these**" it was the sole basis.
- **✓ Roach & Rappaport (1990)** — `roach-rappaport1990_but-they-are-not-thresholds-TLVs.pdf` (+`.txt`). *Am. J.
  Ind. Med.* 17(6):727–753. DOI 10.1002/ajim.4700170607. **[CONFIRMED-primary].** "But They Are Not Thresholds" —
  the TLV documentation does not support the "threshold" claim.

## F. Safety-factor lineage / antecedents (the factor-of-100 trunk and its heirs) — held 2026-06-11
- **✓ Lehman & Fitzhugh (1954)** — `lehman-fitzhugh1954_100-fold-margin-of-safety.pdf` — the **shared trunk**
  (essay-one keystone; register H16). "A good target but not an absolute yardstick."
- **✓ Weil (1972)** — `weil1972_tap_statistics-vs-safety-factors.pdf` — *TAP* 21:454–463; the judgment-over-
  statistics defense ("1/5000 for cancer," set by "scientific judgment"); register H17.
- **✓ Dourson & Stara (1983)** — `dourson-stara1983_uncertainty-safety-factors.pdf` — RfD history; factors "not
  arbitrary as is commonly perceived" (the *defense*); register H18.
- **✓ Barnes & Dourson (1988)** — `barnes-dourson1988_RfD_description-and-use.pdf` — the formal RfD machinery;
  "the original selection of SFs appears to have been **rather arbitrary**" (the *concession*); register H21.
- **✓ Bryan & Shimkin (1943)** — `bryan-shimkin1943_jnci_dose-response-carcinogenic-hydrocarbons.pdf` — the data
  Mantel-Bryan (1961) re-analyzed; register H19.
- **✓ Perera & Boffetta (1988)** — `perera-boffetta1988_jnci_comparing-risks-environmental-carcinogens.pdf` — the
  Ames-HERP critique; register H20.

## G. The no-threshold ROOT (radiation) + the de minimis HEIRS — held 2026-06-11/12, CLOSE-READ 2026-06-13
*The §🔵 "indirect root" rows + the 🔴/🟡 govt/book targets `doc_retrieve` couldn't reach (books, NAP, the full FR
issue), dropped by the operator and read here. Register **H25–H31**. These bracket the lineage at both ends.*
- **✓ Lea (1946/1955)** — `lea1946_actions-of-radiations-on-living-cells_2nd-ed.pdf`; **whole book OCR'd
  2026-06-13** → searchable `…_2nd-ed_OCR.pdf` (PDF/A) + full `.txt` sidecar (24,708 ll.). **Ch. III "The Target
  Theory"** + **Ch. V "Genetical Effects."** The **single-hit / no-threshold ORIGIN**: gene mutation "due to a
  single ionization" → exponential, threshold-free dose-effect; Drosophila yield "**proportional to the dose and
  independent of the intensity**" over a "millionfold" intensity range. Mantel & Bryan cite Lea (ref 12).
  Register **H31**; excerpt `sources/Lea1946_target-theory-single-hit_excerpt.md`. **[CONFIRMED-primary]** (whole book OCR'd).
- **✓ NBS Handbook 59 / NCRP Report 17 (1954)** — `NBS-handbook59_NCRP_permissible-dose-external-ionizing-radiation_1954.pdf`
  (+`.txt`). The **"tolerance dose → permissible dose / acceptable risk"** pivot — no threshold for gene
  mutation, so safety is replaced by a "readily acceptable" risk. Register **H29**; excerpt
  `sources/NBS-Handbook59_1954_acceptable-risk-permissible-dose_excerpt.md`. **[CONFIRMED-primary].**
- **✓ Crump, Hoel, Langley & Peto (1976)** — `crump-hoel-langley-peto1976_cancer-res_fundamental-carcinogenic-processes.pdf`
  (+`.txt`). The **linear-at-low-dose method bridge** (away from Mantel-Bryan probit); footnote 3 = 10⁻⁶ vs 10⁻⁸
  is policy not science. PMID 975067. Register **H26**; excerpt
  `sources/CrumpHoelLangleyPeto1976_linear-at-low-dose_excerpt.md`. **[CONFIRMED-primary].**
- **✓ FDA SOM final rule, 52 FR 49572 (1987)** — `FDA_SOM-rule_52FR49572_1987-12-31.pdf` (+`.txt`; rule at
  txt l. 20693). The **carcinogen-side 10⁻⁶ written into law** via "operational no residue" inside the DES
  Proviso. Register **H25**; excerpt `sources/FDA_SOM-rule_52FR49572_1987_no-residue-10-6_excerpt.md`.
  **[CONFIRMED-primary].**
- **✓ Lowe (1989)** — `lowe1989_rtp_risk-assessment-credibility-fda-perspective.pdf` (+`.txt`). FDA's open
  defense of "a common sense **de minimis interpretation of the 'Delaney clause'**"; narrates *Public Citizen v.
  Young*. *(Originally under a download-tool filename containing "Anna's Archive" — a naming artifact, not the
  source; properly sourced; renamed 2026-06-13.)* Register **H27**; excerpt
  `sources/Lowe1989_FDA-de-minimis-Delaney-defense_excerpt.md`. **[CONFIRMED-primary].**
- **✓ HR 3980 hearing (1961)** — `HR3980-hearing_87thCong_food-additives_feb-mar1961.pdf`
  (=`CHRG-87hhrg66738.pdf`, +`.txt`). **Delaney & Dingell, one room** — extending the 1958 Amendment grace
  period. Register **H28**; excerpt `sources/HR3980-hearing-1961_Delaney-Dingell_excerpt.md`. **[CONFIRMED-primary].**
- **✓ NRC (1996)** — `NRC1996_carcinogens-and-anticarcinogens-human-diet.pdf` (+`.txt`). The **NAS "same
  yardstick"** (mechanistic sameness granted; de minimis conclusion withheld). DOI 10.17226/5150. Register
  **H30**; excerpt `sources/NRC1996_same-yardstick_excerpt.md`. **[CONFIRMED-primary]** (Exec. Summary).

---
*Title- and content-verification log, retrieval routes, and the one caught mis-fetch are recorded in the
session transcript (2026-06-11). The 1995 Threshold of Regulation rule that ties 0.5 ppb to a sub-10⁻⁶ calc is
already held at `sources/FDA_Threshold-of-Regulation_60FR36582_1995.txt`.*