Creation Events, Counted Twice: The Nieh-Yan Charge Law at the Anomaly Ledger and the Quench Bench

Two theorem-grade results of the house Einstein-Cartan program - the legs theorem (the Nieh-Yan density vanishes identically on every static torsion-defect network, paper N2, DOI 10.5281/zenodo.23697937) and the creation-charge law (a defect-creation event carries Nieh-Yan charge exactly Q = b*L, paper N3, DOI 10.5281/zenodo.23698032) - are exported to the two foreign fields where their tolls are comparisons against published numbers, under kill conditions pre-registered in the Transplant Census. ANOMALY LEDGER: producing the axial anomaly coefficient (Delta Q5 = 2 per unit winding per Dirac flavour, exact by the Atiyah-Singer index theorem and Adler-Bardeen) from Q = b*L FAILS and is printed as an obstruction record: (I) species - the law's per-event output is an area, the anomaly's is a pure count, and the only conversion scale (4*l_P^2) would put G inside a coefficient that has none; (II) constancy - b is frame-free at home and the dial table of frame-open choices runs Q/(4 l_P^2) from 0.25 to 7.1 x 10^52, so the target integer is produced at no dial setting and reached only by decree (the double-decree failure of N3). Grade: anomaly coefficient NOT DERIVED; the kill fires. What survives: the event form RENAMES spectral-flow counting (Nielsen-Ninomiya 1983), and the legs theorem CONSTRAINS the geometric page of the anomaly for defect-carried torsion (static networks contribute identically zero). The literature context is printed: the Nieh-Yan term's own anomaly coefficient is unfixed in the mainstream (Chandia-Zanelli 1997: regulator-mass/length-scale dependent; Obukhov-Mielke-Budczies-Hehl 1997 and a 2024 heat-kernel analysis: no Nieh-Yan index or anomaly at all). WITTEN-VENEZIANO: the target is computed in the paper from PDG masses (combination 0.726 GeV^2, requiring chi_YM^(1/4) approx 179 MeV; lattice pure Yang-Mills 191 +/- 5 MeV) and is NOT DERIVED from the charge law, on two independent obstructions - an instanton creates no defect line, so L has no image in that sector, and landing b*L on the implied area (approx 1.20 fm^2) requires dialling both factors to a confinement scale belonging to another Census cell. BENCH FRONT (Kibble-Zurek): the legs distinction (static defects uncharged, creation events charged) is graded RENAMES against the Kibble-Zurek counting discipline (formation-time counting, survivor counting, and the holonomy/divergence species split, with kappa = h/m4 and Phi_0 = h/2e as the bench's holonomy quanta), its falsifier is stated (a change in a static network's divergence-species content with no creation, annihilation, or boundary event), and no bench number is claimed (NOT DERIVED). The bench returns one diagnostic graded CONSTRAINS: b is priced in every laboratory system by a medium (lattice, condensate), and any future home derivation of b must supply a medium-analog structure Einstein-Cartan gravity currently lacks. The vacuum catastrophe is not spent anywhere in the paper and stands NOT SOLVED in the main program.