Normalized gap MAD values confirm that EE is most deformed by the proxy α operator: MAD(EE, α=0.80 vs 1.00) = 6.645, MAD(EE, α=1.00 vs 1.20) = 7.395. TT is intermediate (MAD ≈ 3.6) and TE is lowest (MAD ≈ 2.2). Despite this differential sensitivity, all six channel-type combinations remain locked in Stable Structure throughout — the deformation is measurable but sub-threshold for state transition. The TT gaps ladder shows its minimum ρ̄ exactly at α=1.00 (ρ̄=0.240 vs mean≈0.255), the only CMB group with this property. Two EE levels α values are missing (0.98, 1.01); all other 100 CMB ladders are complete.

The structural sensitivity to α deformation is directly ordered by the gravitational field resolution: Earth's L=720 expansion produces violations at ALL non-nominal α on both coeffmag and degreepower; the Moon (L=300) shows violations on degreepower but only marginal coeffmag violations at α=0.80; Mars (L=85) shows zero coeffmag violations across the full sweep. Richer gravitational fields, encoding more internal harmonic clustering, are more sensitive to α-induced structural deformation. This is the first evidence that the α-sensitivity of a classical field system is a function of the field's structural complexity.

The degree-power table — degree-averaged power spectra of the harmonic coefficients — shows ρ̄ = 0.0177 ± 0.0001 at every α value for all three bodies. This is not a measurement limit — the coeffmag and degreepower ladders from the same runs show large α-sensitivity. The table representation has the same ρ̄ value to four significant figures regardless of whether the gravitational field belongs to Earth, Moon, or Mars, and regardless of α. This is a structural invariant with no equivalent in any other domain tested.

¹⁵⁰Nd sits at the N=90 nuclear shape transition — the boundary between spherical and deformed collective structure. This transitional region produces the most α-sensitive gap ladder in the nuclear corpus. Δρ̄=0.0665 is an order of magnitude larger than most other nuclear groups (typical Δρ̄=0.003–0.009). The ρ̄ minimum is at α=1.02 (ρ̄=0.572), not α=1.00 (ρ̄=0.638). All 17 α points carry violations (min_Ak ≥ 0.998). The MAD structural activity metric (4.70 for ¹⁵²Sm, closely related) confirms the N=90 rare-earth region as the most α-active nuclear zone. This may reflect the high sensitivity of transitional nuclei to small perturbations of the electromagnetic contribution to the level sequence.

At nominal α=1.00, gap ladders carry 2–4× higher structural pressure than levels ladders. This ratio holds across the full α sweep. No α value collapses the two representations to the same pressure class. This confirms that the levels/gaps bifurcation is a geometric invariant of the nuclear ladder architecture — not an artifact of the specific α value or of the nominal fine-structure constant. The bifurcation is an intrinsic property of the gap-structure geometry, invariant to the metric rescaling induced by α deformation. Notable exception: ⁵⁶Fe levels shows the largest Δρ̄ (0.081) of any levels ladder — the iron collective structure has unusual α-sensitivity in the levels representation that does not propagate proportionally to the gaps.

Across 17 α values per domain, 5 domains, and 1,270+ ladder evaluations, the USL admissibility inequality produces no clean violations at any α. The marginal violations observed (geoid: min_Ak=0.619; nuclei: min_Ak=0.991–0.999) are all in the gaps representation and remain far from the clean violation threshold that the cluster adversarial ladders reached (Ak≈0.52, ρ>1 at 14 κ-steps). The USL is non-falsified across a 40% variation of the fine-structure constant — the broadest parameter sweep conducted in this program.

The most striking cross-domain contrast: Earth, Moon, and Mars gravitational fields unanimously identify α=1.00 as the unique ρ̄ minimum across all 12 non-table ladder groups. Nuclear spectra from 14 isotopes identify α=1.00 as the minimum in only 0 of 26 groups. These two classical and quantum domains, both affected by α through electromagnetic contributions, produce diametrically opposite α-response patterns. The geoid operator (α^l harmonic weighting) is perfectly aligned with the real gravitational field structure; the nuclear operator (spin-weighted exponent) captures α-sensitivity but does not reproduce the nominal α as a structural fixed point. This suggests either a deficiency in the nuclear proxy model, or a genuine physical distinction between classical electromagnetic corrections to gravitational harmonics and quantum electromagnetic corrections to nuclear energy levels.

Sodium presents the inverse of the geoid pattern: the nominal fine-structure constant maximises structural pressure in the gap ladder, producing Boundary-Stabilized state (ρ̄=0.605) at α=1.00 while all other α values yield Weak Persistence. This is not a violation — Ak=1.000 throughout — but a qualitative state distinction that makes α=1.00 structurally distinguished in the opposite direction from geoid. The sodium multiplet structure, with its uneven J-resolved fine-structure offsets, creates a gap architecture that is maximally stressed at the physically realised coupling strength. This could reflect the fact that the real sodium spectrum evolved at the real α, producing a level sequence that is maximally spread (highest gap pressure) relative to hypothetical α-deformed variants.

The DESI comoving distance ladder (n=2000) returns ρ̄=0.0261–0.0262 across α=0.80 to 1.20, a range of 0.0001 — indistinguishable from the numerical precision floor. This is the strongest confirmation of Type I metric activity: the cosmological large-scale structure ladder architecture is determined by gravitational clustering geometry, not by electromagnetic coupling, making it genuinely α-invisible at the STRUC-I level. Combined with the knn3 local-structure result (ρ̄=0.533, also at nominal α only), this confirms the local/global structural split while establishing that neither representation is α-sensitive in this domain.

The CMB source-table ladders (raw ℓ-indexed, not value-sorted) show ρ=0.000 from κ=0.01 to κ=0.553, then a step-onset at κ=0.554 in all three channels (TT, TE, EE). This is consistent to three decimal places with κ*≈0.554102 computed from the gap-percolation theory. The α deformation of CMB ladders does not shift this onset (within the current measurement) — the percolation threshold appears to be a structural constant of the CMB acoustic ladder architecture, invariant under the proxy α operator. This is the most precise empirical confirmation of κ* in any observational dataset in the corpus.

The degree-averaged power spectrum table of the gravitational harmonic coefficients produces the same structural pressure reading regardless of body (Earth vs Moon vs Mars — differing in mass, size, density, geological history) and regardless of α (0.80 to 1.20). This is not a floor artifact — the same computation on coeffmag and degreepower ladders from the same run shows large α sensitivity. The ρ̄=0.0177 invariant appears to reflect a universal property of degree-averaged spherical harmonic power spectra at the STRUC-I level. The geoid table occupies the same ρ̄ regime as the CMB source tables (0.013–0.018), suggesting a deep structural equivalence between multipole-averaged CMB power and degree-averaged gravitational harmonic power.

The helium structactive levels ladder shows a clean monotone decrease: ρ̄=0.398 at α=0.80 declining to ρ̄=0.364 at α=1.20, total Δρ̄=0.034. This is not noise — the standard deviation across the sweep is 0.011, and the trend is monotone over 17 points. In the multi-electron atom, increasing α strengthens the electron-electron repulsion term which broadens the energy level distribution, reducing clustering and therefore reducing gap-ladder pressure. This directional response is absent in hydrogen (single electron, no electron-electron repulsion) and reversed/non-monotone in sodium (where the multiplet structure introduces competing effects). The helium result is the cleanest evidence that α acts as a genuine structural operator in multi-electron atoms.

Across all 14 isotopes and all 17 α values, the gap ladder carries consistently higher structural pressure than the levels ladder. The ratio ranges from 1.38 (²⁴Mg) to 4.43 (⁵⁶Fe at nominal α), and these ratios are essentially constant across the α sweep — no α value collapses the two representations to the same regime. This confirms that the levels/gaps bifurcation identified at nominal α is not a property of the specific α value but an intrinsic geometric property of nuclear energy level architecture: the local spacing geometry (gaps) is universally more stressed than the global level distribution (levels), regardless of how strongly the fine-structure constant couples to the system.

The nuclear output pack contains three integrity issues. First, the Mg folder contains ⁵⁶Fe data (correct nucleus, wrong folder label); the Pb folder contains ⁶⁰Ni data. Both sets of results are present and correct — the mislabelling is folder-level only. Second, ¹²⁰Sn gaps contain only 1 α value (α=1.05), making it unusable for sweep analysis. Third, ²⁰⁸Pb is entirely absent from the output — the doubly-magic nucleus with the largest ladder in the corpus (n=608 levels) was not run. Given its structural importance (doubly-magic at Z=82, N=126, predicted highest α-sensitivity in the corpus from spin-weighted MAD=1.78), a complete rerun of ²⁰⁸Pb is required before the nuclear corpus is considered complete.