# The DES "Sensitivity of the Method" rulemakings, 1973 → 1977 → 1979 — where 10⁻⁶ was actually chosen

*The three earlier Federal Register stages of the carcinogenic-animal-drug "no residue" rule, the
contemporaneous record behind the 1987 final (52 FR 49572, register H25). Pulled 2026-06-13 from
`archives.federalregister.gov/issue_slice/…` (operator-supplied exact page ranges). All have text layers.*
**Grade: [CONFIRMED-primary].** Register **H32**. Held locally:
- **38 FR 19226** — proposed rule, **19 Jul 1973** — `papers/lateral/SOM-DES_38FR19226_1973-07-19_proposed.pdf` (+`.txt`)
- **42 FR 10412** — **final** rule, **22 Feb 1977** — `papers/lateral/SOM-DES_42FR10412_1977-02-22_final.pdf` (+`.txt`, slice 10385–10440)
- **44 FR 17070** — reproposed rule, **20 Mar 1979** — `papers/lateral/SOM-DES_44FR17070_1979-03-20_reproposed.pdf` (+`.txt`, slice 17022–17076)

## The finding — the number is dated, and it *moved*
These pages settle the origin question the dossier had carried on Rodricks's firsthand memoir (H1) alone. The
carcinogen "acceptable risk" number entered FDA rulemaking as **Mantel & Bryan's 10⁻⁸**, and FDA **moved it up
two orders of magnitude to 10⁻⁶ in the 1977 final rule** — a deliberate policy choice, documented and rationalized
in the 1979 reproposal. Rodricks attributes the move to GC **Peter Barton Hutt** "after extensive discussions";
the rulemaking record dates its formal adoption to the **Feb 1977 notice** and lays out the reasons.

## 1973 proposal (38 FR 19226) — the number starts at 10⁻⁸ (Mantel-Bryan), called "arbitrary"
*(`SOM-DES_38FR19226_…txt`, ~ll. 205–235)*
- Adopts Mantel-Bryan by name: the required assay sensitivity "will be determined based on the **Mantel-Bryan
  procedure** described in… 'Safety' Testing of Carcinogenic Agents (*J. Natl. Cancer Inst.*, Vol. 27, pp.
  455–470, 1961)."
- The target risk = **1 in 100 million (10⁻⁸):** "If a true exposure of **1/100,000,000** were deemed acceptable
  for an individual on the basis of risk-benefit considerations… The true individual consumer risk would remain
  at **1/100,000,000**." (A 1/10,000,000 = 10⁻⁷ figure appears only as the intermittent-exposure conversion:
  "1/100,000,000 = 1/10,000,000 × 0.10.")
- FDA's own words on the number: "**Absolute safety can never be conclusively demonstrated experimentally.** The
  level defined by the Mantel-Bryan procedure is **an arbitrary but conservative level of maximum exposure…
  (e.g., 1/100,000,000).**"

## 1977 final rule (42 FR 10412) — FDA moves it to 10⁻⁶
*(`SOM-DES_42FR10412_…txt`, ~ll. 1235–1300)*
- The switch, verbatim: "Consequently, the final regulations establish **the maximum risk to be used in the
  Mantel-Bryan calculation as 1 in 1 million.** The following clarifications of the meaning of the 1 in 1 million
  risk level demonstrate why the Commissioner believes that such a risk level can properly be considered **of
  insignificant public health concern.**"
- For scale/rhetoric (a risk-comparison move, like Lowe's "one peanut"): notes "the average risk of fatality by
  motor vehicle accident per year is approximately **1 in 4,000**."

## 1979 reproposal (44 FR 17070) — §"8. Level of risk" narrates and defends the change
*(`SOM-DES_44FR17070_…txt`, ~ll. 2244–2325; two columns de-interleaved)*
- The history, in FDA's voice: "**The 1973 proposal suggested that an acceptable level of risk… could be 1 in
  100 million** over a lifetime. Many comments argued that this level of risk was **unnecessarily conservative**
  … the Commissioner concluded that the **1 in 100 million level of risk was unduly limiting without substantial
  compensation in terms of public health. Consequently, the notice established the maximum risk to be used in the
  Mantel-Bryan calculation as 1 in 1 million.**"
- The selection rule — the "thread-the-needle" criterion, stated outright: "the acceptable risk level should
  (1) **not significantly increase the human cancer risk** and (2), subject to that constraint, **be as high as
  possible in order to permit the use of carcinogenic animal drugs and food additives** as decreed by [Congress]."
- The choice was between round numbers: "**It is difficult to choose between 1 in 1 million and 1 in 10,000 but
  the agency chose the more conservative number** in the general interest of protecting human health." Outside
  suggestions had ranged "**from 1 in 20,000 per lifetime to 1 in 100 million.**"
- And the contrast with the old safety-factor method (the bridge to essay one): "**Historically, safety decisions
  … have been made with the aid of numerical safety factors that do not consider the actual level of risk… an
  absolute safety factor [gives] a 'safe' level. For carcinogens, the Commissioner has concluded that it is
  necessary for the agency squarely to face the level of risk.**"

## Why it matters
- **Answers the origin question with a dated paper trail.** 10⁻⁶ was not derived; it was *chosen* — moved up from
  Mantel's already-"arbitrary" 10⁻⁸ in the **1977 final rule**, because 10⁻⁸ was "unduly limiting" of useful
  products. The criterion was explicitly *both* "not significantly increase risk" *and* "be as high as possible
  in order to permit the use of carcinogenic animal drugs." The round number is a managed compromise, on the
  record, in 1977–79. Corroborates and **dates** Rodricks (H1): Hutt (GC through 1975) drove it in the run-up;
  the Commissioner adopted it in Feb 1977.
- **Flips the last [P-cite] in `10` §3.5/§6/§7** (the earlier SOM rulemaking pages) to **[CONFIRMED-primary].**
- Note the through-line to the first essay: FDA itself frames 10⁻⁶ as the *successor* to the Lehman–Fitzhugh
  "absolute safety factor" method — the two round-number devices, named together in one 1979 paragraph.