FCLT Rival-Constant Catalogue v1: A Fixed Reference Set for Adjudicating Proximity Claims

A proximity screen computes the probability that an arbitrary value falls within δ of a target constant. That answers whether a match is attributable to chance; it does not answer which constant the measurement selects. Where several constants with derivations independent of φ fall inside the same pass band, the screen cannot distinguish them. This record fixes a catalogue of twenty constants, with stated inclusion and exclusion criteria, so that FCLT proximity rows can be re-scored against a reference set timestamped before the re-scoring. Rational convergents of φ (13/8, 21/13, 34/21) are explicitly excluded: they approximate φ by construction and are not independent rivals. Of the nineteen constants other than φ, nine fall inside the 5% band around φ. Median pairwise separation among them is 1.97% against a band half-width of 5.00% — the band is roughly two and a half times wider than the spacing between the constants inside it. Kolmogorov’s inertial-range exponent 5/3 sits 3.01% from φ, so any turbulence measurement reported as a φ match at a 5% threshold is equally a 5/3 match. An adjudication statistic R = δ_rival / δ_claimed is defined, with R ≥ 3 pre-registered as the adjudication threshold. R and significance testing answer different questions and neither substitutes for the other: under Protocol V3.2 no FCLT proximity row attains significance, and R does not change that. This supports a ninth failure class in Protocol V3.2 — rival-constant degeneracy. The catalogue is v1, is a judgement call, and community revision is invited.