Erdős Problem #1032 — We say that a graph is 44-chromatic critical if it has chromatic number 44, and removing any edge decreases the chromatic number to 33.

At least 40 years old · documented by

We say that a graph is 44-chromatic critical if it has chromatic number 44, and removing any edge decreases the chromatic number to 33. Is there, for arbitrarily large nn, a 44-chromatic critical graph on nn vertices with minimum degree ≫n\gg n?

References

Progress summary

Refreshed
Claimed progress

The problem remains open: known constructions give only sublinear minimum degree, while a new unverified argument gives a linear upper bound.

The problem asks whether arbitrarily large edge-critical graphs requiring four colours can have minimum degree proportional to their number of vertices. Erdős reportedly posed it more than twenty years before the historical record.

Known results

  • Simonovits and Toft independently constructed 44-critical graphs with minimum degree growing as n1/3n^{1/3}.
  • Luo, Ma, and Yang obtained the upper bound δ(G)<0.328n\delta(G)<0.328n.
  • Dirac constructed 66-critical graphs with minimum degree greater than n/2n/2; the analogous 55-critical question remains open.

Recent claimed linear upper bound

GPT-5.5 Pro + harness was credited with the claimed inequality 8e(G)/n2+δ(G)/n≤3/2+o(1)8e(G)/n^2+\delta(G)/n\leq 3/2+o(1), implying δ(G)≤(3/10+o(1))n\delta(G)\leq(3/10+o(1))n. This is a partial result, not a resolution, and remains unverified independently.

Current status (as of June 2026): The existence of 44-critical graphs with minimum degree Ω(n)\Omega(n) remains open; n1/3n^{1/3} constructions and a claimed, unverified 0.3n0.3n-type upper bound are known.

Sources

Solutions 0

No solutions have been posted yet.