Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Statement
Here is a finite simple graph on vertices containing no cycle of length four as a subgraph, and is its maximum degree.
Lemma (p. 188, unnumbered, quoted). "Let be a quadrilateral-free graph on vertices. If the maximal degree of , , satisfies , then ."
The Lemma carries no parity or prime-power hypothesis on ; enters only through the number of vertices.
Refinement (p. 189, unnumbered). After the Lemma's proof, closing Section 3, the paper records that the same proof gives more: if satisfies the Lemma's hypotheses and for some , then
Since , the right side equals , so the refinement improves on the Lemma's bound once (an observation of this page; the paper states the refinement without further comment).
Source. Z. Füredi, Graphs without Quadrilaterals, J. Combin. Theory Ser. B 34 (1983), 187-190: Section 2, the unnumbered Lemma on p. 188; its proof is Section 3 on p. 189, which closes with the refinement. The edition read is identified on the source card.
Read depth. Claims checked: the statement and the refinement were read clause by clause on the printed pages. The proof was followed in outline only; no independent proof review is recorded.
Proof pointer
Section 3, p. 189. Fix a vertex of maximum degree . Since two vertices have at most one common neighbour, each other vertex has all but at most one of its neighbours outside that vertex's neighbourhood, and counting pairs outside the neighbourhood bounds a sum of binomial coefficients of the degrees. Assuming more than edges, Jensen's inequality turns this into the polynomial inequality (3), which the two inequalities (4) and (5), valid for , contradict.
Dependencies
None beyond the definitions; the proof is self-contained.
Used by
The Theorem and the Proposition of the same paper, which treat the remaining case using that is even.
Bears on
- Problem 765: an upper bound on at the orders , valid only for graphs of maximum degree at least . It is an ingredient of the paper's exact values at those orders and does not by itself give the asymptotic formula the problem asks for.