All three hydrogen QM-I representations (preprocessed, gap structure, spectrum) return HARD with GR ∈ [0.814, 0.934] — the only QM-I family that fails the FULL class. The hydrogenic n⁻³ level spacing creates extreme gap heterogeneity (gaps grow without bound with quantum number) that no encoding can regularize. This is a representation-invariant HARD result: Theorem 1 is triggered regardless of how the hydrogen ladder is constructed. By contrast, Zeeman H (magnetically split, finer-scale gaps) returns GIANT near the corner.

Iron Zeeman has tailDom = 0.967 — far from the helium zone (tailDom > 0.995). The helium separator F = GR − tailDom = 0 gives F = +0.030 (FULL side) for Fe-ZEE, but Fe-ZEE is GIANT. This is not an anomaly: it confirms exactly what the local theory of the main manuscript requires. The separator F=0 is a local geometric object, valid only in the narrow (tailDom, GR) zone explored by the helium corpus. Outside that zone, different chart geometry applies. Fe-ZEE is the first corpus system to demonstrate this locality explicitly. Within the narrow zone (tailDom>0.995): 13/13 correct separation.

Light atoms (H, He, Na) have Zeeman tailDom > 0.999 (corner zone C); heavier/more complex atoms fall into the interior (Ag: 0.998, Fe: 0.967) or TAIL (Au: 0.998 but GR<1). The Fe outlier (Z=26 but tailDom=0.967) reflects iron's dense d-electron fine-structure: many moderate-scale magnetic gaps dilute the outlier fraction, unlike lighter elements with sparser Zeeman splitting. Au (Z=79, simpler s-valence structure) stays in the corner zone by tailDom but reaches TAIL by GR. This is the first cross-system structural trend in the corpus.