The nucleosome's "~1.65 turns" — is it fixed, whose frame is it in, and what shape is it?¶

Research spike (2026-07-19; concertmaster dispatch). Scoping/derivation only — no code shipped, no rc, no ADR. FORM-matching only: biology's X has the same cascade-shape as srmech's Y; this does not validate/extend the framework and biology is not superseded ([[user_stance_cascade_matching_substrate_blind_form_not_identity]], [[feedback_no_lineage_claims_in_notebook]]). Provisional throughout. Companion to chromatin_histone_structural_machinery_findings.md (row 1 / gap G3) and subharmonic_chirality_carrier_findings.md. Generating script for every number below: nucleosome_turn_asymmetry_frame_spike.py ([[feedback_computational_provenance_discipline]]) — exact integer/Class-N rationals, Class-K pin-slot for sign, no abs(), no float in any load-bearing result.

0. The methodological gate — resolved FIRST, and it fires¶

The gate fires harder than the dispatch assumed. Do not fit the decimal.

Attested candidate constants against the target, and against the real spread:

candidate	value	dev. from 1.65
φ = (1+√5)/2	1.61803	0.03197
5/3	1.66667	0.01667
147/89	1.65169	0.00169
28/17 (= 14/8.5)	1.64706	0.00294
√e	1.64872	0.00128

Worst deviation among all candidates = 0.05. Against that:

canonical-particle band alone: 1.65 – 1.70 → width 0.05
attested physical spread across particle classes: 1.20 – 1.90 → width 0.70 (14× larger)
the sign is not fixed: handedness flips between −0.80 ± 0.05 and +0.86 ± 0.39 turns at a barrier of only 2.3 ± 0.4 k_BT [PMC4623960]

Every candidate sits inside the rounding band of one crystal structure. The target discriminates nothing. Any fit is unfalsifiable numerology. [NULL — gate fires; no fitting performed below.]

The φ-proximity is therefore explicitly not evidence of a phyllotaxis/optimal-packing shape. It is 2% off a number whose real physical range is 58% wide.

1. Q1 — Is it even fixed? NO. It genuinely deviates. This is the spike's first-order finding.¶

The literature spread is not measurement scatter. It decomposes cleanly:

Convention/construct artifact (real scatter, resolved): - 146 vs 147 bp — settled at 147. The dyad lines up with a base pair, so the native count is odd; Luger 1997 used a 146 bp palindrome on the pre-structural assumption of an even count and the particle absorbed the 1 bp deficit by stretching [PMC4378457, McGinty & Tan 2014]. - "1.65" vs "~1.7" — same measurement, different rounding. - "1.75" — no OA source found. Treat as spurious. [NULL]

Real physical variance (the finding):

particle	wrapped bp	turns	source
H2A.B nucleosome	103	1.2	PMC7780145
H3–H4 octasome	~120	1.5	PMC9659345
canonical NCP	145–147	1.65–1.7	PMC4378457, PMC7780145
chromatosome (+H1)	166–167	1.9	PMC7801413

Plus: salt-dependent unwrapping 7 ± 2 bp → 22 ± 5 bp across physiological ionic strength [PMC8129070]; in vivo range ~100–170 bp [PMC8129070]; ΔLk per nucleosome varies −1.4 to −0.9 as a systematic function of nucleosome spacing [PMC5659657]; and the handedness inverts, with wrapping orientation set by the pre-assembly supercoiling state of the DNA, i.e. not uniquely determined by the octamer [PMC7959483].

Farr et al. 2021 name it directly: nucleosomes are "a dynamic family of particles", "highly dynamic and structurally irregular entities" rather than "static building blocks" [PMC8129070].

And the primary source said so in 1997. Luger et al.'s own abstract closes:

"The lack of uniformity between multiple histone/DNA-binding sites causes the DNA to deviate from ideal superhelix geometry." [PMID 9305837, abstract fetched]

Q1 VERDICT: "the exact rational" was the wrong question. There is no fixed value to find. The user's reframe #2 is correct on the evidence — the object deviates, and the deviation is a function of variant, salt, sequence, spacing, and supercoiling state. This supersedes the framing of row 1 / G3 in chromatin_histone_structural_machinery_findings.md, which recorded "1.65 turns" as a fixed quantum.

Premise corrections logged (three prompt premises did not survive): H1 does not take the wrap to an integer 2 turns (attested 1.9); Klug & Lutter 1981 report 10.0 bp/turn, not 10.6 (the 10.6 is Rhodes & Klug 1980 for DNA on a flat surface, identified with the solution value); the ~76 bp unwrapping is the high-force transition, not the first [PMC9388122].

2. Q2 — Whose frame? The k=2 frame-split reading is CONFIRMED at the geometry level and REFUTED at the "integer invariant" level.¶

2.1 Attestation status: now attested (it was not in-tree before this spike)¶

Lk = Tw + Wr and the linking-number paradox were absent from chromatin_histone_structural_machinery_findings.md. Both are now OA-attested:

Lk is an integer topological invariant; Tw and Wr are geometric and trade off continuously — Benham 2024, NAR 52(1):22–48, DOI 10.1093/nar/gkad1092 (fully OA):

"L is an integer, and has a fixed value so long as both DNA strands remain covalently closed." "Both W and T can change continuously as a DNA domain fluctuates… But in a ccDNA topoisomer their sum, which is L, must remain a fixed integer."
Dennis & Hannay 2005, arXiv:math-ph/0503012v2 (full text extracted): "this topological invariant is the sum of two other terms … which individually depend on geometry rather than topology."

2.2 The frame-split reading — SUPPORTED, and stronger than posed¶

The literature does not merely treat 1.65 turns as writhe-side; it publishes the explicit conversion Wr = n(1 − sin δ) [PMC6162219, Segura et al. 2018]. So "turns" and "writhe" are not even the same geometric quantity — there are two geometric stages before any topology:

   1.65 turns  --(pitch-angle δ≈4°)-->  Wr ≈ −1.53  --(+ΔTw)-->  ΔLk
   [frame-dependent]                    [frame-dependent]        [measured]

Reproduced exactly (Class-N rationals, srmech sin_series_truncate): −1.65 × (1 − sin 4°) = −1.5349 vs published −1.53. ✓ Sensitivity: δ=3° → −1.564; δ=5° → −1.506.

2.3 The refutation: ΔLk per nucleosome is NOT the integer invariant¶

This is where the dispatch hypothesis breaks, and it matters:

The integer invariance predicate is closure of the whole domain — "so long as both DNA strands remain covalently closed" [Benham 2024]. It lives on the whole minichromosome, not per nucleosome.
Segura 2018 measures ΔLk = −7.07 for the whole minichromosome and −1.26 per nucleosome — both non-integers, obtained by subtracting the means of two topoisomer distributions, which are Boltzmann-populated because "the energy difference between the Lk topoisomers is less than the thermal energy" [PMC6162219].
Călugăreanu–White–Fuller does not apply to open segments at all. Sierzega, Wereszczynski & Prior 2021 [PMC7811023]: "the linking … is not an invariant for open-ended ribbon structures"; "there is no evidence that the equality in (1) holds if the ribbon is not closed"; and artificial closure "will generally contribute to both the writhing and the linking of the composite."

Q2 VERDICT: ~1.65 turns is a frame-dependent geometric half of a k=2 pair — attested, twice over. But the other half is not an integer invariant; it is a real-valued, ensemble-averaged, thermodynamic local linking difference (Fuller 1978's reference-ribbon construction is what lets a per-nucleosome number be extracted from a globally-closed molecule at all [PMC392823, abstract]). Both members of the pair are on the geometry/statistics side. The integer sits above both, at the whole-domain scale, where no per-nucleosome number exists. The honest object is therefore not an exact rational and not an integer — it is a distribution.

2.4 The frame-ledger reconciliation (exact)¶

Every published "disagreement" closes its own ledger — because Lk = Tw + Wr is a theorem:

account	regime	Wr	ΔTw	ΔLk	closes?
classical textbook	ΔLk assumed −1.0	−1.70	+0.70	−1.00	✓ resid 0
Segura 2018	ΔLk measured −1.26	−1.46	+0.20	−1.26	✓ resid 0
Nikitina 2017	ΔTw held at 0	−1.70	0	−1.70	✓ resid 0

The three OA accounts differ only in which member is held/assumed. But one gap is not a frame choice: the classical postulates ΔLk = −1.00 while Segura measures −1.26 — a 0.26 gap between an assumption and a measurement. The frame reading does not dissolve it. Separately, Segura's own geometric Wr(−1.53) + ΔTw(+0.20) = −1.33 vs measured −1.26, a 0.07 gap they attribute to DNA breathing — i.e. Q1's variance re-entering the topology.

3. Q3 — What is the shape? (broad enumeration; two survive, three fail)¶

3.0 Direct answer to the user's reframe #1 — and its honest limit¶

"What discrete op⊗operand⊗responsion structure, at what coherency, has this value as its projection — with a FORMULA, not a fit."

Answer, for the twist term: the discrete substrate quantity is the integer 14 — the octamer's minor-groove contact count [ATTESTED, PMC4512544]. The coherency is exact commensuration, where the DNA's helical repeat divides the wrap into exactly one turn per anchor:

   coherency condition:   N / h = 14      ⟺   h = N/14 = 147/14 = 21/2 = 10.5 bp/turn
   the formula:           ΔØ = N · (1/hs − 1/h0)        [detuning from that coherency]
   at hs = 51/5, h0 = 21/2, N = 147:  ΔØ = 7/17  exactly

The continuous quantities — 10.2 bp/turn, ΔØ, ΔTw — are projections of the integer 14 under a detuning. (14 anchors bound 13 unit intervals plus a half-turn overhang at each end: 13×10.5 + 2×5.25 = 147 = 14×10.5. Both accountings agree.)

The limit, stated plainly: this derives ΔØ. It does NOT derive 1.65. The wrap count is a separate geometric fact (superhelix pitch and radius), and per Q1 there is no fixed 1.65 to derive — so the correct conclusion is not "we haven't found the formula yet" but "the quantity the reframe asked about turns out not to be a constant." The formula exists for the member that is structurally determined.

#	candidate shape	mechanism	independent observable predicted	attestation	verdict
S1	Two-periodicity beat / moiré	surface hₛ ≈ 10.2 vs solution h₀ ≈ 10.5 bp/turn; the detuning integrated over the wrap is the twist term	ΔTw should track (1/hₛ − 1/h₀)·N across any surface-wrapped DNA, not just nucleosomes	numbers ATTESTED [PMC6162219]; derivation is ours	SURVIVES, but under-determined — §3.1
S2	ℤ/14 contact-lattice commensuration	14 minor-groove-inward anchors, one arginine each, at SHL −6.5…+6.5	sub/super-nucleosomal particles wrap k × 10.5 bp for their own contact count k; and a computable ΔLk per k	contacts ATTESTED [PMC4512544]	DEGRADED (2026-07-19) — fails a k-independent 3.49σ test and cannot represent the observed bistability — §3.2.1
S3	structure-blind 10.5 bp quantisation	"all wrap lengths are near multiples of 10.5"	variant bp counts cluster at multiples of the spacing	—	NULL — §3.2
S4	φ / phyllotaxis / optimal divergence	golden-angle packing	a 137.5° divergence between successive contacts	—	NULL — observed advance is 42.4°/contact; and §0 kills the fit
S5	Kuramoto / Arnold-tongue mode-locking	phase-lock of two coupled periodicities with a critical coupling	a coupling-strength threshold below which the lock breaks; hysteresis	—	NOT TESTED — fermata F-c

3.1 S1 — the beat, exact, and its hazard¶

Exact Class-N arithmetic with hₛ = 51/5, h₀ = 21/2, N = 147:

  1/hs − 1/h0 = 1/357  turns per bp
  ΔØ = 147 × 1/357   = 441/1071 = 7/17 = 0.411765 turns   [147 = 3·7²; 1071 = 3²·7·17]

ΔØ = 7/17 exactly, reproducing Segura's published "ΔØ ≈ +0.4". A second, independent verification run converged on the same value from the other direction (147/10.2 = 14.41 vs 147/10.5 = 14.00) — convergence, not dissonance.

This is a VERIFICATION of Segura et al.'s published arithmetic, not a discovery. Both halves of their decomposition now reproduce from published geometry alone: ΔØ = 7/17 ≈ +0.412, and −1.65(1 − sin 4°) = −1.5349 vs their Wr = −1.53; their ΔTw ≈ +0.2 closes as +0.412 − 0.19 (STw) = +0.222. What the framework adds is the reading — that the term is a commensuration detuning off an integer lattice — not the numbers. [[feedback_no_lineage_claims_in_notebook]].

And at the solution periodicity the commensuration is exact:

  N / h0 = 147 / (21/2) = 14 exactly   ← 14 helical turns == 14 contacts, one turn per contact
  N / hs = 147 / (51/5) = 245/17 = 14.4118
  detuning                            = 7/17

So the twist term is the detuning of the DNA from an exact 14-fold commensuration with the octamer's contact lattice. That is a genuine closed-form reading, and the two-periodicity beat is real.

The hazard, which is load-bearing. ΔØ is a difference of reciprocals and is therefore hypersensitive to inputs that the OA literature does not pin. Across the attested periodicity values:

hₛ	h₀	ΔØ exact	ΔØ
10.0 (Klug & Lutter 1981)	10.5	7/10	0.700
10.1 (Chandrasekhar 2024)	10.5	56/101	0.554
10.2 (Segura 2018)	10.5	7/17	0.412
10.4 (Bishop 2008)	10.5	7/52	0.135

Span = 0.632 turns — comparable to the entire physical spread of ΔLk (−0.9 to −1.5, width 0.6) that the term is invoked to explain. So the periodicity-difference "resolution" of the paradox is not quantitatively constrained by the attested data. This independently explains why the three OA accounts disagree on mechanism (core overtwist vs. linker geometry vs. more-negative-ΔLk): the beat term is free enough to absorb any of them. [ANOMALY — logged §5.]

3.2 S2/S3 — the contact lattice; a null and a survivor¶

ATTESTED [Hodges et al. 2015, Genetics, PMC4512544]:

"These interactions occur primarily at 14 locations in the nucleosome structure where the DNA minor groove faces the histone octamer [superhelical locations (SHL) −6.5 to 6.5]." … "At each of these locations, an arginine side chain extends into the DNA minor groove."

S3 (structure-blind) is NULL. Testing "is bp/spacing near an integer" across 10 attested particles, against a Uniform[0, s/2] null (mean s/4, var s²/48; z² reported exactly):

spacing	mean \|resid\|	null	z²	\|z\|	verdict
10.5 bp	2.300	2.625	338/735	0.68σ	NULL — better than random but far under 2σ
10.0 bp	2.800	2.500	54/125	0.66σ	NULL — worse than random

[NULL — structure-blind quantisation has no support.]

S2 (structure-aware) survives. Using each particle's own contact count k and bp = k × 10.5:

particle	contacts k	predicted bp	observed	resid
octamer NCP	14	147.0	147	0.0
hexasome (−1 H2A–H2B)	11	115.5	110–120	−0.5
chromatosome (+H1)	16	168.0	166–167	−1.0
H2A.B	10 (predicted)	105.0	103	−2.0
H3–H4 octasome	11	115.5	~120	+4.5
tetrasome (H3–H4)₂	6	63.0	~70	+7.0 ✗

⅚ within half a contact-spacing. Circularity caveat (load-bearing): k is usually read off the same structures the bp count comes from. The reading is non-circular only where k is fixed independently by which histone fold is deleted — hexasome, tetrasome, chromatosome. The one genuine open prediction is H2A.B ⇒ k = 10, which would need an independent contact count to test.

Independent-observable the surviving pair jointly predicts (the falsifiable core): the contact count and the surface periodicity are not independent facts — 14 anchors distributed over 147 bp is the constraint forcing hₛ ≈ 10.2 rather than 10.5. So k → hₛ → ΔØ → ΔTw → ΔLk is one causal chain, and the prediction is: a particle with a different contact count k should show a different surface periodicity, hence a different ΔTw, hence a shifted ΔLk — with the shift computable in closed form. For a hexasome (k=11): predicted hₛ = 115.5/11.29… — testable against measured ΔLk for hexasomes. Not tested here (no OA hexasome ΔLk located). [FERMATA F-a.]

3.2.1 S2 DOWNGRADED — the first per-particle ΔLk at k ≠ 14 (amendment, 2026-07-19)¶

Verdict change: S2 goes from "survives with a caveat" to DEGRADED. It is NOT recorded as having survived a test. Source of the datum and the σ computation: the open-experiments spike and its committed provenance script. Structural checks below independently re-derived here.

The datum [ATTESTED-OA, CC BY] — Vlijm R, Lee M, Ordu O, et al. (2015), PLoS One 10(10):e0141267, DOI 10.1371/journal.pone.0141267, PMC4623960: the tetrasome ΔLk is a bistable pair, flipping between −0.80 ± 0.05 and +0.86 ± 0.39 turns, barrier 2.3 ± 0.4 k_BT. This is the first per-particle ΔLk at k ≠ 14 in our ledger — exactly the class of test F-a asks for.

S2 fails it three ways, in increasing severity:

A 3.49σ miss using the OBSERVED wrap — and this is k-independent. Because the test uses the measured N rather than S2's own bp = k × 10.5, the k=6-vs-7 ambiguity cannot rescue it. (4.74σ using the S2-predicted wrap.)
The residual SIGN FLIPS between particles — +0.053 at canonical (N=147) vs −0.175 / −0.237 at tetrasome (N=70). Re-derived independently: a correction term c·N needs c = +0.000361 for canonical but c = −0.002500 / −0.003386 for the tetrasome branches; and a term monotone in N cannot change sign between two positive N. No law linear in N fits both.
Decisive — an arity failure, not a fit failure. S2 is a function k → ΔLk: one input, one output. The observed particle occupies two states at the same k, separated by 1.66 turns and thermally interconverting. S2 cannot represent this particle at all. No re-parameterisation fixes arity.

Why it doesn't cleanly die — and why that is the indictment, not a reprieve. The σ miss can be absorbed by re-choosing unattested auxiliary scalings. That freedom is exactly anomaly A1 — the beat term is under-determined to the width of the quantity it explains, so it can absorb almost any residual handed to it. Therefore:

A hypothesis that cannot be falsified is not passing a test when it survives one.

This REINFORCES A1; it does not resolve S2. S2's survival here is a symptom of the under-determination, not evidence for the shape.

Diagnosis in in-tree vocabulary — naming the failure, NOT repairing it. The two branches are near-equal in magnitude (0.80 vs 0.86, differing by 0.06) and opposite in sign — the "±-pair at equal magnitude, differing only by orientation" of subharmonic_chirality_carrier_findings.md §1–2. So the precise defect is that S2 is a magnitude-only law with no chirality degree of freedom ("a lone theta is a lone chirality"). Flagged explicitly: this is a diagnosis, not a consolation result and not a rescue. S2 stays DEGRADED. Whether a ±-pair-valued successor is worth building is a conductor question, not a claim here.

F-a remains OPEN. The tetrasome tests S2 but does not close F-a: the hexasome (k=11) is still the cleanest test — no bistability confound. The open-experiments spike could not locate hexasome ΔLk (deep null: four search families plus raw-byte grep across six papers).

3.3 Frame both-directions check (`[[feedback_always_check_both_directions_including_time]]`)¶

Every candidate above assumes the octamer frame (n turns of DNA about a fixed protein). Reciprocal frame — hold Wr, solve n = Wr/(1 − sin δ):

given	n (turns)	Class-N anchor
ΔWr = −1.46 (Segura implied)	1.5695	11/7
ΔLk = −1.26 (Segura measured)	1.3545	23/17
ΔLk = −1.70 (Nikitina, ΔTw=0)	1.8275	53/29
ΔLk = −1.00 (classical)	1.0750	29/27

The reciprocal-frame values span 1.08 – 1.83 — reinforcing §0: which number you call "the turns" depends entirely on which member you held.

4. The k=3 reading — `Lk ⊗ Tw ⊗ Wr` (conductor's mid-task angle)¶

Epistemic ceiling, stated first and hard. Lk = Tw + Wr is an established theorem of differential geometry (Călugăreanu 1959/61, White 1969, Fuller 1971). Everything below is our framework's reading of a structure that already exists and is already documented. Nothing here is a discovery, and the mathematics is not "secretly k=3." Form-matching only.

4.1 The type-asymmetry is real, and the three-way scope distinction is ATTESTED¶

All three integral forms attested from Dennis & Hannay 2005 (arXiv:math-ph/0503012v2, full text):

member	integral form	domain	depends on	value type
Tw	`(1/2π) ∮_A ds (t × u)·u̇`	1-fold, one curve	axis + framing	real
Wr	`(1/4π) ∮_A ds ∮_A ds′ …`	2-fold, A×A (self)	axis only	real
Lk	`(1/4π) ∮_A ds ∮_B ds′ …`	2-fold, A×B (two curves)	neither — invariant	integer

"Tw is local in the sense that it is an integral of quantities defined only by s on the curve, and clearly depends on the choice of framing (ribbon)." "writhe Wr equals the sum of signed nonlocal crossings … twice the average of self-crossings of the axis curve with itself" "The crossings between the two edge curves naturally fall into two types: 'local,' which will be associated with Tw, and 'nonlocal,' which will be associated with Wr."

Note the three-way locality structure is the source's own organizing principle, not our imposition — Dennis & Hannay build their proof on the local/nonlocal crossing split.

Where the integer comes from — ATTESTED:

"The domain of integration in equation (3.2), A × B, is the cross chord manifold … topologically equivalent to the torus … with the torus 'wrapping around' the sphere an integer number of times (the integer arises since the cross chord manifold has no boundary, and the mapping is smooth)"

[SYNTHESIS — ours, not attested] The integral-arity ladder framing (1-fold / 2-fold-self / 2-fold-pairwise), and the inference that the invariant sits on Lk and not Wr despite both being Gauss double integrals because the self-domain A×A carries a diagonal singularity at s = s′ while A×B is closed and boundaryless. Dennis & Hannay state each ingredient; the ladder reading is ours. Flagged so it is not mistaken for a cited result.

4.2 Why this argues natively triadic rather than 2+1 — the differing-bipartition test¶

The strongest honest support is not "there are three terms" (trivial) but that different criteria pick different odd-ones-out:

by value type (integer vs real): Lk stands out
by integral arity (1-fold vs 2-fold): Tw stands out
by dependency set: all three differ — Tw depends on axis+framing, Wr on axis only, Lk on neither

If the object were a k=2 pair with a labelled third, every criterion would yield the same bipartition. It does not. That is a checkable statement, and it is the only part of the k=3 reading that earns its keep on evidence rather than aesthetics.

4.3 Candidate assignment — weighed, not assumed¶

The conductor proposed Tw = op, Wr = operand, Lk = responsion. This aligns with srmech's own shipped responsion schema, which keys on (operator, carrier) → responsion{answers_with, status} — i.e. an acting member, a carried member, and an answering correspondence that verifies (srmech.amsc.responsion_schema, 23 edges, rc225). On that shape:

Tw = op — the local action applied pointwise along the curve ✓ (attested local)
Wr = operand — the global embedding being acted on ✓ (attested axis-only, self-referential)
Lk = responsion — the pairwise correspondence between the two ribbon edges, invariant, integer ✓ (attested two-curve; and "stored relationship between two strands" is on the nose for srmech)

Alternatives weighed and rejected as worse fits, not as impossible: (a) Wr = op — rejected, writhe is a property of the axis, not an action; (b) Lk = op — rejected, Lk is not an action and is invariant under the deformations op would perform; © Tw = responsion — rejected, Tw is framing-dependent and so cannot be the verifying member. But note the algebra is symmetric — Tw = Lk − Wr and Wr = Lk − Tw are equally valid rearrangements. What breaks the symmetry is the type/locality structure of §4.1, not the algebra. Absent that table, the assignment would be unmotivated.

4.4 The three "centrisms" — 2 attested-direct + 1 derived¶

regime	physical realization	attestation	status
hold Lk → Tw/Wr trade	ccDNA topoisomer; single-molecule torsion on chromatin fibers	Benham 2024; Kaczmarczyk 2020 [PMC6949304]: "At fixed linking number, this must be compensated by increased twist in the DNA handles"	ATTESTED, canonical
hold Tw → Lk/Wr co-vary	ΔTw=0 idealization; torsionally-relaxed / nicked DNA	Nikitina 2017: "if the DNA Twist is not changed (ΔTw = 0)"; Corless & Gilbert 2016 [PMC5153829]: "Most of the linker DNA in eukaryotes is torsionally relaxed"	ATTESTED (caveat: nicking destroys closure, so Lk ceases to be defined, not merely varies)
hold Wr → Lk/Tw co-vary	DNA axis path pinned by the octamer's 14 contacts	DERIVED from attested premises — §4.4.0. (The surface-adapted SLk formalism is separately OA-attested, §4.4.1, but is a different decomposition and is not the basis.)	DERIVED (mathematically entailed; physically approximate)

k=3 VERDICT (revised 2026-07-19): the composition is 2 attested-direct + 1 DERIVED. Stated in exactly those terms so this is never later read as three citations. A derived result is first-class here, not a lesser tier — see the methodology note in §4.4.0.

4.4.0 Regime (iii), DERIVED — the derivation, shown so it can be checked¶

This is OUR derivation, not a citation. It is written out step-by-step precisely so a reader can check it against premises they can fetch themselves, rather than take our word for it.

Premises (all attested, all openly fetchable):

P1 — Wr = W[A], a functional of the AXIS CURVE ALONE. Dennis & Hannay (arXiv:math-ph/0503012v2): both writhe integrations run over A; Benham 2024 NAR: "W is a geometric parameter determined by the shape of the central axis curve C."
P2 — Lk = Tw + Wr, for a closed ribbon. Benham 2024 NAR; Dennis & Hannay.
P3 — the octamer's 14 contacts pin the DNA axis path. Hodges et al. 2015 [PMC4512544]: 14 minor-groove-inward locations, an arginine inserted at each. Corless & Gilbert [PMC5153829]: "each nucleosome in the genome constrains a single under-wound supercoil."

Derivation:

D1. From P1, Wr is a function of A. Therefore fixing A fixes Wr. Direction check (the step most likely to be got backwards): {A fixed} ⇒ {Wr fixed} is sufficient, not necessary — distinct axis curves can share a writhe. Only sufficiency is needed to exhibit a realization, so the direction used here is the valid one.
D2. From P2 with Wr held constant: ΔLk = ΔTw + 0, i.e. ΔLk = ΔTw exactly — Lk and Tw co-vary one-for-one. That is regime (iii).
D3. D1 + D2 use no biology whatsoever. The regime is mathematically well-posed on the theorem plus the axis-only dependency alone. This is stronger than the "physical realization" framing this note carried before the re-examination.
D4. From P3, the nucleosome physically realizes the constraint — approximately. Q1's breathing/unwrapping/variance is exactly the size of the approximation, and is not hidden here.
Inherited condition: closure of the ribbon (P2) — the same condition regimes (i) and (ii) inherit. No extra assumption is smuggled in for (iii).

Entailment verdict, stated honestly: the mathematical claim is entailed (D1–D3, airtight). The physical claim is approximate (D4), bounded by Q1. Regime (iii) is therefore well-posed.

Consistency test — apply the regime to Segura's own data and check whether the residual is a known term or an unexplained one:

   hold Wr = −1.53  (octamer geometry)
   measured ΔLk     = −1.26
   ⇒ ΔTw implied by regime (iii) = ΔLk − Wr = 27/100 = +0.270
     Segura's stated ΔTw          =            +0.200
     residual                     =             0.070

The residual 0.07 is not new and not unexplained — it is the same breathing gap already isolated in §2.4. The regime reproduces the known ledger and its residual lands on a term the literature already names. Consistent.

Methodology — standing practice from 2026-07-19, not a one-off. A paywalled result is not a dead end: derive it from open premises and show the steps. The rule against unquotable sources exists for open quotability, not cost — a source no reader can open makes them trust us instead of check us. A derivation from premises anyone can fetch is therefore more re-verifiable than a citation nobody can read, and re-verifiability is what attestation is for. A result labelled DERIVED is first-class. Guard-rail: deriving is not permission to assume — if the premises do not entail the conclusion, say so and leave the gap open.

4.4.1 The SLk decomposition — what the F-b hunt did and did not deliver¶

FOUND [ATTESTED-OA, CC BY-NC] — Chen B, Xiao Y, Liu C, Li C, Leng F (2010), "DNA linking number change induced by sequence-specific DNA-binding proteins," Nucleic Acids Research 38(11):3643–3654, DOI 10.1093/nar/gkq078, PMC2887952 (Europe PMC reports isOpenAccess: Y, license: cc by-nc):

"White et al. also showed that ΔLk can be described by two geometrical terms: the surface linking number (SLk) and the winding number (ϕ). In this case, ΔLk = ΔSLk + Δϕ."

and, applied to the paradox itself:

"It has been known for a long time that 146 bp of DNA wrap around the histone octamer core 1.8 turns in a left-handed superhelix. In this case, ΔSLk = −1.8. Initially, it was mis-expected that ΔLk = ΔSLk, which resulted in the 'linking number paradox' … ΔLk = ΔSLk + Δϕ, where Δϕ = −0.8 and significantly compensated the ΔSLk to yield ΔLk = −1."

Corroboration [ATTESTED-OA, CC BY] — Segura 2018 [PMC6162219], already in our ledger, turns out to carry the White-1988 attribution explicitly and states the twist-side form: ΔTw = ΔØ + ΔSTw, where "the winding number (Ø) depends on the helical repeat of DNA at the nucleosome surface (hs), and the surface twist (STw) is a correction function that accounts for the curved path of DNA." Our existing attestation was stronger than recorded.

NULL — the load-bearing gap. No OA source states that SLk is independent of the helical repeat / determined by the surface alone, nor "hold the surface fixed ⇒ SLk fixed." A Europe PMC full-text search for the exact phrase "surface linking number" returns 10 records in the entire indexed literature, of which exactly one is in the OA subset — the formalism lives almost entirely in paywalled 1988–1994 Science/JMB/Springer literature. Confirmed misses: Benham 2024 NAR — zero occurrences of SLk/STw/White (the highest-probability hit, a clean miss); Nikitina 2017 cites without restating ("according to the surface linking theory (24)") — rejected on exactly the cite-vs-restate line; Prunell 1998 and White/Gallo/ Bauer 1989 NAR are free but scanned images, no text layer; Leng 2016 Biophys Rev restates the equation but is free-to-read, NOT OA-licensed (isOpenAccess: N) — a weaker tier, not used; Swigon 2009 paywalled; arXiv full-text search returned 0 entries.

4.5 Falsifiability — stated plainly, including what CANNOT falsify¶

What CANNOT falsify it: the closure condition Lk − Tw − Wr = 0. It is a theorem; it closes with residual exactly 0 in all three published accounts (§2.4) and could not have done otherwise. Any test built on the closure is vacuous. The dispatch's proposed falsifier ("the three descriptions must give mutually consistent accounts of the same data") is therefore not a test — it is guaranteed. Reported as such.

What WOULD falsify it, and the current standing: 1. If Tw were also a Gauss double integral (same kind as Wr and Lk) → 3-scope claim collapses. → Survives: attested 1-fold. 2. If any two members shared an identical dependency set → not natively triadic, it is 2+1. → Survives: §4.1 table, all three differ. 3. If any of the three hold-one-fixed regimes were ill-posed or physically unrealizable. → SURVIVES (2026-07-19). Regimes (i) and (ii) attested-direct; regime (iii) derived and well-posed (§4.4.0), with the derivation shown so it can be checked. Note the honest asymmetry: (i) and (ii) have directly attested physical instances; (iii) has an entailed mathematical well-posedness plus an approximate physical realization. That is a difference in kind of evidence, not a shortfall — but it is stated rather than smoothed over. 4. If the differing-bipartition test (§4.2) failed — i.e. some criterion made the same bipartition canonical across all three axes. → Survives on the three criteria tested.

4.6 Numerology hazard — the codon-radix k=3 link, checked and rejected¶

Per instruction, tested rather than assumed. Two independent hazards, both cleared:

k=3 (triality) vs k=3 (codon radix): both are 3. That is a coincidence of the smallest non-trivial error-correcting arity until a shape correspondence is shown. Lk/Tw/Wr is a one-invariant-two-projections triad; the codon radix is a 4³=64 alphabet cardinality. No shape match demonstrated. Do not link. [FLAGGED, not used as evidence.]
14 contacts vs the framework's 14 A–N classes: the totals coincide; the partitions do not. A–N is 1+3+7+3 (sorted [1,3,3,7]); the nucleosome contact inventory is 3+3+3+3+2 by histone-fold dimer (sorted [2,3,3,3,3]). Cascade-matching compares shape, not cardinality, and there is no part larger than 3 on the biology side. [NULL — hazard explicitly cleared, and worth keeping cleared.]

5. Diagrams¶

The Tw/Wr frame-split, and where the integer actually lives:

graph TD
  subgraph GEOM["frame-DEPENDENT (geometry) — where ~1.65 lives"]
    n["n ≈ 1.65 turns<br/>(superhelical wrap, octamer frame)"]
    Wr["Wr ≈ −1.53<br/>2-fold ∮∮ over A×A (SELF)<br/>depends on AXIS only"]
    Tw["ΔTw ≈ +0.20<br/>1-fold ∮ over A (LOCAL)<br/>depends on AXIS + FRAMING"]
    n -- "Wr = n(1 − sin δ), δ≈4°" --> Wr
  end
  subgraph STAT["ensemble / thermodynamic — NOT an invariant"]
    dLk["ΔLk ≈ −1.26 per nucleosome<br/>mean of a Boltzmann topoisomer distribution<br/>varies −0.9 … −1.5 with spacing"]
  end
  subgraph TOPO["frame-INVARIANT (topology) — the integer"]
    Lk["Lk of the WHOLE closed minichromosome<br/>2-fold ∮∮ over A×B (PAIRWISE)<br/>INTEGER — domain is a boundaryless torus"]
  end
  Wr --> dLk
  Tw --> dLk
  dLk -- "only defined because the WHOLE domain is closed" --> Lk
  classDef geom fill:#eef,stroke:#446;
  classDef topo fill:#efe,stroke:#464;
  classDef stat fill:#ffe,stroke:#a84;
  class n,Wr,Tw geom;
  class Lk topo;
  class dLk stat;

The two-periodicity beat (why the octamer's 14-fold lattice detunes the DNA):

   contact lattice (octamer):   |....|....|....|....|....|....|....|....|....|....|....|....|....|....|
                                 1    2    3    4    5    6    7    8    9   10   11   12   13   14
                                SHL −6.5 ......................... 0 ......................... +6.5
                                 └── 14 anchors, one arginine into the minor groove at each ──┘

   DNA at SOLUTION periodicity h0 = 10.5 bp/turn :  147 / 10.5 = 14.000 turns   ← EXACTLY commensurate
   DNA at SURFACE  periodicity hs = 10.2 bp/turn :  147 / 10.2 = 14.412 turns   ← detuned

                                          detuning = 7/17 = 0.412 turns  ==  ΔØ
                                                    │
                                  this beat IS the twist term in the paradox

6. Anomalies¶

A1 — The beat term is under-determined to the width of the thing it explains. ΔØ spans 0.135–0.700 across the attested periodicity inputs (span 0.632), while the whole physical ΔLk spread is 0.6. The periodicity-difference resolution of the linking-number paradox is therefore not pinned by OA data. Investigation: exact reciprocal-difference arithmetic, 4 attested hₛ × 2 attested h₀. Verdict: real; independently explains the live three-way mechanism disagreement in the OA literature (Segura core-overtwist vs Nikitina linker-geometry vs classical). Next: an OA source that pins hₛ with error bars would collapse it; none located.

A2 — Assumption-vs-measurement gap of 0.26 turns. The classical ΔLk = −1.00 is a postulate; Segura's −1.26 is a measurement [PMC6162219]. Frame-reading does not dissolve this — it is a factual disagreement, and the textbook value is the weaker of the two. Next: conductor decision on whether any in-tree row may cite −1.0.

A9 — a retrieval failure against ourselves: we already held the answer and filed it under the wrong question. The tetrasome ΔLk pair (−0.80 / +0.86, PMC4623960) that downgraded S2 was already in-tree — quoted in §0 of this very note — but filed as a handedness anecdote ("the sign is not fixed"), not as a per-particle ΔLk at k ≠ 14. F-a was asking for exactly that, and a separate spike had to re-find it. Investigation: cross-check of §0 against the open-experiments spike's find. Verdict: real, and ours. The datum was never missing; the index was wrong — it was catalogued by the property that first caught our attention rather than by the quantity it measures. Next: when a number is recorded, record what quantity it is and what tests it could serve, not only the point it was cited to make. Note this is the same failure mode as A5 one level up: A5 found the literature conflating four quantities as bare numbers; A9 finds us filing one quantity under the wrong question. The MPM ledger indexes provenance well and queryability badly.

A4 — the Chen 2010 SLk ledger does NOT close arithmetically as printed. The quoted sentence gives ΔSLk = −1.8, Δϕ = −0.8, ΔLk = −1. But −1.8 + (−0.8) = −2.6 ≠ −1. The word "compensated" requires opposite signs, and Δϕ = +0.8 closes exactly: −1.8 + 0.8 = −1.0. Investigation: exact rational differencing (provenance script, block A4). Verdict: a sign error in the printed text or in the extraction — unresolved either way. The magnitude 0.8 is usable; the sign as printed is not. Do not rest anything on the −0.8 figure. Next: re-fetch the published PDF/typeset version to see whether the minus is in the paper or in the XML extraction. Math doesn't lie — flagged rather than quietly corrected.

A5 — four DIFFERENT quantities are reported in the literature as bare numbers near 1–2. The spread "1.65 / 1.7 / 1.8 / 1.9" is not four measurements of one thing:

value	context	quantity
1.2 / 1.5 / 1.65 / 1.7 / 1.9	variant wraps	superhelical turns
1.8	Chen 2010, "146 bp … 1.8 turns"; ΔSLk = −1.8	surface linking / wrapping number
−1.53	Wr = n(1 − sin δ)	writhe
−1.26 / −1.00	per-nucleosome (measured / postulated)	linking difference

The 1.8 is not a sixth measurement of the wrap — it is a different quantity. So part of the apparent "disagreement about 1.65" in the literature is not disagreement at all; it is the same frame confusion this spike is about, appearing in the source literature itself. This independently reinforces §0 (the target does not discriminate) and §2 (the honest object is frame-specified or it is nothing). Verdict: real; next: any in-tree row quoting a bare "turns" number must name which quantity it is.

A3 — Search-summary text did not survive fetch. A search result attributed "~10.2 bp periodicity … 14 independent sites" to Jin, Rube & Song 2016 NAR; on fetch the sentences were not in the paper (it discusses 10.5 bp and never mentions 14 sites). Logged as a live instance of the hallucination vector MPM exists to catch. Do not cite that pairing.

7. Fermatas (conductor decisions — this pass is NOT authorized to decide)¶

F-a (the real test). — STILL OPEN 2026-07-19; partially answered, not closed. The k→hₛ→ΔØ→ΔTw→ΔLk chain predicts a computable ΔLk shift per contact count. The tetrasome ΔLk has now been found (Vlijm 2015, PMC4623960) and S2 fails against it (§3.2.1) — but the tetrasome does not close F-a, because its bistability confounds the test that F-a was designed to run. The hexasome (k=11) remains the cleanest instrument — single-valued, one dimer removed, no chirality confound. Hexasome ΔLk is a deep null: four search families plus raw-byte grep across six papers, nothing. Dispatch a dedicated hunt, or accept the null and let §3.2.1 stand as the verdict on S2?
F-b (regime iii). — CHASED 2026-07-19; PARTIAL FIND + a residual NULL. An OA restatement of the SLk decomposition was found (Chen et al. 2010, PMC2887952, CC BY-NC — §4.4.1), and our existing Segura attestation turned out to carry the White-1988 attribution already. But regime (iii) is still not closed: SLk is a different decomposition, and the invariance property the regime needs has no OA restatement (10 records corpus-wide; 1 in the OA subset). k=3 stays at ⅔ + 1-by-inference — the standard was not weakened to close it. Anomaly A4 (the Chen ledger does not close as printed) is open and should be resolved before that source is used for anything numeric.
F-h (abstract-tier attestation). — RESOLVED 2026-07-19. My own proposal was WITHDRAWN, correctly. I had raised whether the free White 1988 abstract was admissible. It is not, and the reasoning corrects the premise I raised it on: the rule's rationale is OPEN QUOTABILITY, not cost. A source no reader can open makes them trust us instead of check us. So: no abstract-tier, no free-to-read tier, no laundering of unquotable authority into the record. White/Cozzarelli/Bauer 1988 is a pointer only, never a basis. The gap it appeared to leave was not a gap — the answer was already in hand as a derivation (§4.4.0), and a derivation from open premises is more re-verifiable than an unreadable citation. Standing practice recorded in §4.4.0.
F-c (S5 untested). Kuramoto/Arnold-tongue mode-locking was enumerated but not tested. srmech ships cascade.kuramoto_step with Sakaguchi-α + directed adjacency — the instrument exists. Worth a spike, given §3.1 shows the system is a two-periodicity detuning problem?
F-d (notebook placement). — CONFIRMED + EXECUTED 2026-07-19. Q1 supersedes the "1.65 = fixed quantum" framing of row 1 / G3 in chromatin_histone_structural_machinery_findings.md. That note is now amended: row 1 re-keyed to the integer contact count k; new §2.1 carries the four-particle-class table and the four tooling consequences; G3 amended (keying on k strengthens it — hexasome / tetrasome / chromatosome become first-class rather than exceptions); both compaction ladders updated; and the "147 bp = exactly 49 codons" weak anomaly in that note's fermata F-a is withdrawn (it rested on 147 being a constant). Per the split rule, the tooling/structural consequence stays srmech-side there; the ontological point (deviation-as-content; the honest object being a distribution) is cross-ref'd to MFO only and not restated in the srmech note.
F-e (provenance script). The dispatch said "no code"; [[feedback_computational_provenance_discipline]] requires generating code for load-bearing numerics. Script written as a research artifact, not package code. Keep, or strip?
F-f (running-notes convention). — CONFIRMED 2026-07-19. srmech_mpm_notes.ndjson kept as the direct peer of docs/antikythera-maths/research-mfo/mfo_mpm_notes.ndjson (same phase: "concertmaster_dispatch" schema). That peering is now the convention for srmech-side running notes.
F-g (user framing, 2026-07-19) — MFO-SIDE, RECORDED NOT ADJUDICATED. The user's reading: DNA's shape carries both continuous and discrete domain mathematics, and this is a physical instantiation of the asymmetric universal resonator shape MFO finds in the cosmos. This is an ontological claim and MFO-side material per [[feedback_mfo_vs_srmech_notebook_split_rule]]; this note records it as a cross-ref and does not adjudicate it. Two guard-rails attach, and they are load-bearing:
It must not retro-justify S1 or S2. Those two survive on their own attested footing (§3.1, §3.2) or not at all. A resonator framing is not evidence for a commensuration derivation.
It does not soften anomaly A1. The beat term remains under-determined to the width of the very quantity it explains (span 0.632 vs 0.6) — §6. An ontological reading that makes the deviation meaningful does not make the arithmetic constrained. Those are independent questions. The observation the framing rests on is in-tree and attested here: the object genuinely carries a discrete integer invariant (k = 14 contacts; whole-domain Lk) and continuous frame-dependent members (Tw, Wr, the wrap) — §4.1. What that means is MFO's to decide.

8. Sources (attestation status)¶

Attested OA, full text fetched this spike: Segura et al. 2018 Nat Commun 9:3989 PMC6162219 · Nikitina et al. 2017 Sci Adv PMC5659657 · Sierzega et al. 2021 Sci Rep 11:1527 PMC7811023 · Dennis & Hannay 2005 arXiv:math-ph/0503012v2 · Benham 2024 NAR 52(1):22–48 DOI 10.1093/nar/gkad1092 · Hodges et al. 2015 Genetics PMC4512544 · McGinty & Tan 2014 Chem Rev PMC4378457 · Corless & Gilbert 2016 Biophys Rev PMC5153829 · Kaczmarczyk et al. 2020 Nat Commun PMC6949304 · Farr et al. 2021 Nat Commun 12:2883 PMC8129070 · Zhou K, Gaullier & Luger 2018 PMC7386248 · Zhou M et al. 2021 EMBO J PMC7780145 · Wang et al. 2021 Sci Rep 11:380 PMC7801413 · Vlijm et al. 2015 PLoS One PMC4623960 · Vlijm et al. 2017 Nanoscale PMC7959483 · Nozawa et al. 2022 PNAS PMC9659345 · Klug & Lutter 1981 NAR 9(17):4267 PMC327434 · Díaz-Celis et al. 2022 PNAS PMC9388122 · Talbert & Henikoff 2021 J Cell Sci PMC8015243 · Winger et al. 2018 eLife DOI 10.7554/eLife.34100 · Rudnizky et al. 2017 Protein Sci PMC5477540.

Abstract-only (body not fetchable — flagged, not rested on): Luger et al. 1997 Nature 389:251 PMID 9305837 (the 1.65 figure is not in the abstract; attested secondhand via PMC4378457) · Prunell 1998 Biophys J PMC1299595 · Fuller 1971 PNAS PMC389050 · Fuller 1978 PNAS PMC392823 · Vasudevan, Chua & Davey 2010 JMB PMID 20800598.

Added by the F-b hunt (2026-07-19): Chen B, Xiao Y, Liu C, Li C & Leng F (2010) NAR 38(11):3643–3654 DOI 10.1093/nar/gkq078 PMC2887952 — ATTESTED OA (CC BY-NC); restates ΔLk = ΔSLk + Δϕ and applies it to the paradox. ⚠ its printed ledger does not close — see anomaly A4; do not use its Δϕ sign.

REJECTED (not openly quotable ⇒ not attestation, and NOT our basis): White, Cozzarelli & Bauer 1988 Science 241:323 — POINTER ONLY: the surface-invariance result is independently stated there (paywalled, not openly quotable, not our basis). Our basis for regime (iii) is the derivation in §4.4.0 from open premises. Its free abstract is not admitted — see F-h; the rule's rationale is open quotability, not cost. · White, Gallo & Bauer 1989 JMB 207:193 (Elsevier) — same status, pointer only · Swigon 2009 IMA vol. 150 (Springer) · Rhodes & Klug 1980 Nature 286:573 · Călugăreanu 1959/61 and White 1969 originals — no OA copy located.

Free-to-read but NOT OA-licensed (weaker tier — not used): Leng F (2016) Biophys Rev 8(3):197–207 PMC5425792 — restates ΔLk = ΔSLk + Δϕ, but Europe PMC reports isOpenAccess: N with no license field. Recorded to show the tier distinction was applied, not to rest a claim on.

Rejected on cite-vs-restate (fetched, but only CITES the formalism): Nikitina et al. 2017 PMC5659657 ("according to the surface linking theory (24)") · Prunell 1998 PMC1299595 and White/Gallo/Bauer 1989 NAR PMC318199 — free but scanned page images, no extractable text layer.

NULL / could not attest: "1.75 turns" (no source) · the 5-vs-6 bp contact-spacing irregularity · a uniform bp spacing between adjacent contacts (partial only) · hexasome ΔLk.

Cross-links: chromatin_histone_structural_machinery_findings.md (row 1, G3) · subharmonic_chirality_carrier_findings.md · ../../antikythera-maths/subharmonic_chirality_collapse_stub.md · MFO §XIV.8 (F1179–F1186, op⊗operand⊗responsion) · srmech.amsc.responsion_schema · [[user_stance_k2_compare_is_frame_relative_asymmetric_pair]] · [[feedback_cad_ban_is_gpu_numerical_not_closedform_physical]] (closed-form topology in scope).