One-Octonion Brane-Bulk Framework - Paper CCCXC: SPARC Rotation Curves Without Dark Matter, and the 1 Mpc Lensing-Inferred Curves on the Sump Premise: The T1 Law Against Smith et al. (2026), a Measured Central Deficit, and the Retraction of the Inner-Region Claims

One-Octonion Brane-Bulk Framework, Paper CCCXC. Smith et al. (arXiv:2610.33157) fit 147 late-type SPARC rotation curves with cold, warm and fuzzy dark-matter halos and with MOND, and extend the mean curves to 1 Mpc with KiDS weak lensing. This paper refits the same 147 galaxies, reproduced to the galaxy, with the framework's T1 sump law g = gbar ν(gbar/a0), a0 = cH0/2π = 1.042×10−10 m s−2, and with every inner-region extension the corpus has proposed. Three results. (i) With distance and inclination held at their SPARC values the T1 law is MOND: median reduced χ2 3.37 against 0.93 for the feedback halo DC14, and a0 = 1.042 is indistinguishable from 1.2 (3.37 vs 3.30), the difference being absorbed by a 15% shift in the stellar mass-to-light ratio. With distance and inclination freed under the SPARC priors, the T1 law reaches 2.18 with three parameters, within 0.25 of the warm-dark-matter halos with five (1.93, 1.96) and behind NFW (1.49) and DC14 (1.16); with the κ-window amplitude added (four parameters) 1.62. Freeing a0 per galaxy buys only 2.18 → 2.01, and per-galaxy BIC finds the median galaxy indifferent between a fixed and a free a0 (ΔBIC +0.9), with a quarter of galaxies still strongly preferring their own scale. The 1 Mpc flatness of the lensing-inferred circular velocities, which no halo reproduces, is the deep limit of the T1 law, v4 = G Mb a0, a phantom mass linear in r with no virial radius; the framework has no derived light-deflection law, so this leg tests the Paper XXVIII premise that the sump apparent mass lenses as ordinary mass together with the T1 law, and is not an independent prediction of weak lensing. (ii) The inner-region extensions of Papers XXXIV and LI are not supported: the three-term law with positive amplitudes gives 3.05, the zero-parameter form JBH = e·a0 of Paper 0 §9 gives 6.69, and the CS-124 figure χ2/dof = 0.6 is not reproduced; Prediction 13 is withdrawn. The CCCXXV closure and gate, a Gudermannian interpolation, and a uniformly hyperbolic coherence cell all leave the median at 3.2–3.5. (iii) The residual is measured rather than modelled: a central cusp of missing acceleration, 27% in the innermost tenth of rT2* = √(GMbar/a0) falling to zero by 0.75 rT2* (21% with distance and inclination freed), present in two thirds of galaxies and confined to star-dominated regions. The canonical window −κeff = tanh2((rT2*/r)2) (Paper CXXIII) bounds it, and a curvature radius diverging toward a rupture point reproduces its shape (universal median 3.08), but the fitted rupture radius is 1.4 kpc, about 0.55 Rdisk, with slope +0.12 ± 0.10 against baryonic mass: it does not scale with rT2*. The author's reading is a bulk-perpendicular turn of the sump current at the drain, which projects the in-brane acceleration by cos θ(r); the data then require θ ≈ 40° at 0.05 rT2* on a mass-independent kiloparsec scale. Registered test P110: resolved inner kinematics (BIG-SPARC) decide between that and inner-disc kinematics. Files: the paper (PDF) and a reproducibility archive with every fit script, the per-galaxy tables, the 127-check gate script and its recorded run, the paper source, the seven working notes, and the two-round cross-model (Sakana/Fugu) audit with raw replies and text snapshots. SPARC data: Lelli, McGaugh & Schombert (2016), astroweb.cwru.edu/SPARC/. Anchor DOI: 10.5281/zenodo.19120873. Community: one-octonion-brane-bulk. Author: Bharathi Dasan Jagadeesan, M.D., University of Minnesota. ORCID: 0000-0002-1143-941X.