Coloring and immersion

High ★★★ Graph Theory » Coloring » Vertex coloring accessible to undergrads

Status partial high confidence

The conjecture that every graph $G$ with $\chi(G) \ge t$ immerses $K_t$ remains open in full generality. Post-posting, DeVos–Kawarabayashi–Mohar–Okamura (2010) confirmed it for $t \le 7$ via the stronger fact that minimum degree $t-1$ suffices; Gauthier–Le–Wollan (2017) showed chromatic number at least $3.54t+4$ guarantees a $K_t$-immersion; and Fox–Pach–Suk (2025) proved the conjecture for graphs on at most $(1.64-o(1))t$ vertices.

Cited literature (3)

  • Matt DeVos, Ken-ichi Kawarabayashi, Bojan Mohar, Haruko Okamura · Ars Mathematica Contemporanea

    Proves the conjecture for $t \le 7$ via the stronger result that every graph with minimum degree $t-1$ immerses $K_t$.

  • Gregory Gauthier, Tien-Nam Le, Paul Wollan · arXiv preprint · arXiv:1703.08235

    Shows that every graph with chromatic number at least $3.54t+4$ immerses $K_t$, a linear sufficient condition weaker than the conjectured threshold of $t$.

  • Jacob Fox, Janos Pach, Andrew Suk · arXiv preprint · arXiv:2510.05893

    Proves the Abu-Khzam–Langston/Lescure–Meyniel conjecture for graphs on at most $(1.64-o(1))k$ vertices, where $k$ is the chromatic number.

Reviewer notes. The conjecture was independently formulated by Lescure and Meyniel in 1989, predating Abu-Khzam–Langston. No journal publication was located for the Gauthier–Le–Wollan arXiv preprint (1703.08235). The Fox–Pach–Suk (2025) paper targets Albertson's crossing-number conjecture but proves the immersion conjecture as an intermediate result for graphs with at most $(1.64-o(1))k$ vertices; it is an arXiv preprint as of the review date. The conjecture itself remains fully open for all $t$.

Auto-reviewed 2026-05-08 with claude-sonnet-4-6 (web search enabled) · 152s.

Conjecture. For every positive integer $ t $ , every (loopless) graph $ G $ with $ \chi(G) \ge t $ immerses $ K_t $ .
Keywords: coloring · complete graph · immersion

Discussion

Let $ G $ be a graph and let $ uv, vw \in E(G) $ . The operation of deleting the edges $ uv $ and $ vw $ and then adding a new edge between $ v $ and $ w $ is called a split . We say that a graph $ G $ immerses a graph $ H $ if a graph isomorphic to $ H $ may be obtained from $ G $ by repeatedly making splits and deleting vertices and edges. The graph containment relations of minor and topological minor have been thoroughly studied with respect to graph coloring. In particular, there are two famous conjectures: Hajos conjectured that every graph with chromatic number $ \ge t $ contains a subdivision of the complete graph $ K_t $ as a subgraph. Hadwiger conjectured that every graph with chromatic number $ \ge t $ contains $ K_t $ as a minor. While Hajos' Conjecture is false for $ t \ge 8 $ (indeed it is actually false on average), Hadwiger's Conjecture remains open, and is one of the outstanding problems in Graph Theory. On the other hand, the relationship between graph coloring and immersions seems to have been largely ignored until Abu-Khzam and Langston made the above conjecture. In addition to formulating this conjecture, they proved it for $ t \le 4 $ and showed that a minimal counterexample to it must be 4-connected and $ t $ -edge-connected. Recently, DeVos, Kawarabayashi, Mohar, and Okamura have resolved the conjecture for $ t \le 7 $ .

Bibliography