# README for the independent verifier — AP49 The Hold, freeze bundle of 6 September 2026

**The 420 Code · Studio G · issued on the author’s word, 6 September 2026**

This bundle freezes Artist’s Proof 49, *The Hold*, FINAL v1.0, with everything its provenance points to. Nothing in it is to be edited; anything found wrong is answered by a dated note beside the file.

## What is in it
- `AP49_The_Hold_FINAL_v1_0.md` — the text of record; `.docx` and `.pdf` are renderings in the house dress (AP43’s).
- `inputs/` — AP30 *The Resistance* as published (the paper whose switch KS-30.3 this proof pays); the corpus’s pre-registration of 2 August 2026 (`prereg_2026-08-02-mp-me-alpha3.md`, digest 6ea78dac…, as retrieved from the420code.org/prereg on 6 September), which AP49 supersedes by reference.
- `instruments/` — `PROV_HOLD_1_2026-09-06.md` (the disposition of every review finding, v0.1 → v0.6; roles only); `DATED_NOTE_AP30_2026-09-06.md` (AP30’s CODATA label; the sixteen’s reading; KS-30.3 → PAID); `PREREG_2026-09-06-mp-me-closed.md` (the superseding pre-registration entry: D_closed beside D, what changed, the sign of c₃, the kill conditions, the version chain).
- `lineage/` — every retained draft, v0.1 through v0.6.
- `SHA256SUMS.txt` — digests of every file above, this README included.

## What to verify
1. `sha256sum -c SHA256SUMS.txt`.
2. The pre-registration of 2 August against the site: digest 6ea78dac4e47a4c9….
3. The paper’s numbers from its inputs (CODATA 2022: α = 7.2973525643(11) × 10⁻³, m_p/mₑ = 1836.152673426(32)): r = 16α/1836 = 6.3593 × 10⁻⁵; the tail 21α·r/(1 − r) = 9.7459657 × 10⁻⁶; D_closed = 1836.152673444951; residual −1.895 × 10⁻⁸ = −0.59σ; the pin 15.97 ± 0.05; c₃ = 21·16²/1836² = 1.595 × 10⁻³; the third order 6.197 × 10⁻¹⁰, the fourth 3.941 × 10⁻¹⁴; the corpses (7.1 × 10⁻⁸ at 2.8σ; 2.8 × 10⁻⁷ at 9.5σ; 3.3 × 10⁻⁵ at ~10³σ; 1.5 × 10⁻¹⁰ at 0.6σ); the leaked base 1836.152673408; the kill at 3σ once the bar is 3.4 ppt, 5σ at 2.1 ppt; twelve sigma between the bases at 1.7 ppt.
4. The H₂⁺ value at its source: Alighanbari, Schenkel, Korobov, Schiller, Nature 644, 69 (2025), doi 10.1038/s41586-025-09306-2.
5. That the third order was given before the arithmetic — PROV-HOLD.1 records the order; the author’s two answers of 6 September are quoted in §3 and §4.

## What this paper claims, in one line
m_p/mₑ = 1836.152673445, the series closed, the third order +6.2 × 10⁻¹⁰ with its sign stated in advance; it dies at 3σ once the measurement reaches 3.4 parts per trillion with its central value unchanged, and it has no negative term with which to repair itself.

*Nature agrees, or nature doesn’t.*