s1104 — neutrino Δm² ratio: reading A (X_ν literal) vs reading B (resolvent rung d=2)
  NuFIT 6.1 IC24 NO (suite-carried): Δm²₂₁ = 7.537e-05 · Δm²₃₁ = 2.511e-03 · ratio = 0.03002
  m₂/m₃ FLOOR (m₁=0, NO) = √ratio = 0.17325 ± 0.00233 (calibration-grade σ, 1.35%)

(1) THE TWO READINGS ARE ONE FAMILY
  [PASS] reading A ratio φ⁻¹/√5 = 0.27639  (= (1−φ⁻²)/(φ²−φ⁻²)) — A712 value
  [PASS] reading A's m₂/m₃ = λ₂/λ₃ = φ⁻¹ = 0.61803 = R₁ = 1/(φ²−1) = 0.61803  ⇒ A IS THE d=1 RUNG
      factor vs measured: A/meas = 9.21× (refuted, unchanged) — the suite's d-assignment puts ν at d=2, not d=1

(2) READING B SCORED — R₂ against the m₁=0 floor
  R₂ = 1/(φ⁴−1) = 0.17082 · floor 0.17325 · deviation -1.40% · z = -1.04σ (calibration-grade)
  [PASS] R₂ lies BELOW the m₁=0 floor (so m₁>0 can only WORSEN the tension)
  [PASS] |z| < 2σ at m₁=0 — NOT refuted; an EDGE prediction
  R₂² = 0.02918 vs measured ratio 0.03002 (-2.79%) — the number Paper 2 never printed
  m₁ needed to push R₂ to −2σ: 1.4 meV · to −3σ: 2.0 meV  ⇒ B requires m₁ ≈ 0 (< 1 meV)
  [PASS] Inverted ordering: m₂/m₃ → 1 ≫ R₂ — B PREDICTS NORMAL ORDERING (as does A)

(3) COROLLARIES READING B FORCES (m₁ = 0, NO) — all falsifiable, none printed in the suite
  m₁ = 0 · m₂ = 8.68 meV · m₃ = 50.11 meV · Σm_ν = 58.8 meV
  m_β (KATRIN observable, suite angles s12²=0.3044, s13²=0.02145) = 8.74 meV
  m_ββ = 0 under the suite's DIRAC-LOADED treatment (Appendix X :933); a Majorana branch would re-open it
  cosmology comparator (CALIBRATION-GRADE, not source-lined here): Σm_ν < ~64–72 meV (95%, DESI+CMB 2024–25)
  [PASS] Σm_ν prediction sits BELOW ~72 meV (2024 bound) — alive
  ⚠ margin to the tighter ~64 meV bound is 5.2 meV: the next cosmology release is a live test of B

(4) LOOK-ELSEWHERE — what menu was R₂ drawn from
  menu: 20 numbers in [9.71e-06, 0.62] · tolerance ±2% · hits on m₂/m₃ floor: ['R_2']
  log-space coverage ≤ 7.2%  ⇒ a chance hit at this tolerance has p ≲ 0.07; R₂ is the ONLY hit
  [PASS] exactly one menu member within 2% of the floor
  X_ν seesaw-power class m_i ∝ λ_i^k (k=1..4): m₂/m₃ = 0.6180, 0.3820, 0.2361, 0.1459 — none within 5% of the floor
  [PASS] no X_ν power k=1..4 within 5% of the floor

VERDICT
  The suite prints reading A as its one falsified row and carries reading B one paper away with no
  comparison. They are the d=1 and d=2 rungs of the same R_d family. Scored honestly, B is an EDGE
  prediction: m₂/m₃ = R₂ sits -1.40% (-1.04σ, calibration-grade) below the m₁=0 floor, so it
  stands ONLY with m₁ ≈ 0 and NO — and then forces Σm_ν ≈ 59 meV, m_β ≈ 8.7 meV, m_ββ = 0 (Dirac-loaded).
  NOT a hit, NOT refuted, NOT tier-moving: the d=2 assignment for ν is 'structural' in print (Rev29) and
  PARKED by A869; whether d=2 is DERIVED or SELECTED is the PI ruling this kernel asks for. Route S293.1.

s1104: ALL CHECKS PASS
