# T737 argument selection The ordered integer parameters a,b are joint coordinates of one Fourier index: omega=(-1+i*sqrt(3))/2, lambda=1+2*omega, m=a+b*omega=lambda^3*mu. Both parameters have category **arithmetic_structure**. They are neither two independently sampled real inputs nor selectors between mathematical objects. The natural domain is the nonzero Fourier lattice lambda^(-3)*Z[omega]; Patterson's support theorem DR (5.7)--(5.8) selects its nonzero coefficients. The declaration describes the natural lattice, not just this run's radius. The finite selection includes every nonzero supported m with 00 or (2*a-b=0 and b>0), ordered by norm,a,b. The sign representative uses proved evenness and keeps all other unit phases. Proven zero support and the constant term are explained, not filled as rows. No quotient by complex conjugation or multiplication by omega is used. The bound 1000 is a noncanonical computational cutoff invariant under numeral notation. It keeps each direct Gauss sum below 1000 residue classes while covering split and inert primes, composites, all four formula lines, successive ramified levels, and nonunit cube factors, including shared primes in c,d. Useful breadth was selected before extra digits or a deeper squarefree slice. There are no continuous couplings and no exploratory reference-sample groups. The 511 expected rows are a consequence of this joint region, not a quota. Enumeration and full decomposition checks are executable in generate.py and independent.py; actual results accompany this document. Completeness is no: the natural lattice is infinite and this run has a finite radius. Sources inspected: Dunn--Radziwill arXiv:2109.07463v3 §1.2 and §5.1 (5.6)--(5.8); David--de Faveri--Dunn--Stucky arXiv:2410.03048v2 (6.8) for the integral sublattice. The archived canonical bundle manifest is ../20261117T112154Z-build-skills/manifest.json, bundle SHA-256 984965681e2f8925c511a121c118efcde59581c7b0d906cac8b641bf6925e55a. This is the builder's rationale, not the independent argument-selection verdict.