For G_0 a minimum-order 5-chromatic maximal planar graph and any
4-coloring of G_0 - uv, the endpoints u, v must share a color, and the
color classes pairing that color with each of two other colors must
each induce a u-v path. The Kempe-chain parts follow from a standard
swap-on-component contradiction against the shared-color claim.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Defines D(G) as the family of single-edge-deletion spanning subgraphs
of a maximal planar graph G, and shows that when G_0 is a minimum-order
5-chromatic maximal planar graph every member of D(G_0) is 4-colorable,
via a coloring pulled back from the smaller minor G_0/uv.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds a transitional section reframing the frequency results: the
relevant class is not all maximal planar graphs but those that resist
Kempe-style reductions, where flip-asymmetry's exclusion may have
real bite. Sets up subsequent development of additional necessary
properties of a minimum-order 5-chromatic counterexample.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The unrestricted census suggested flip-symmetry already excludes a
vanishing fraction of maximal planar graphs; this commit re-runs the
same enumeration over the minimum-degree-5 subclass (where any
minimum-order 5-chromatic counterexample must live) to check whether
the restriction tightens the bound. It does not: the density decays
to zero there as well, only at a gentler geometric rate (~0.63 per
step instead of ~0.5).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
didericis