Wiki
Wiki

Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.

Updated


Statement

As printed on p. 78 (PDF p. 4 of the typescript scan, page image): "Let GG be the skeleton of a cube. Simonovits and I proved [9]

f(n;G)<cn8/5(7)f(n;G)<cn^{8/5} \tag{7}

We could not decide whether (7) is best possible."

The skeleton of a cube is the graph Q3Q_3 of its vertices and edges; f(n;G)f(n;G) is the smallest number of edges forcing GG, so (7) is ex(n;Q3)<cn8/5\mathrm{ex}(n;Q_3)<cn^{8/5}. The paper's [9] is the 1970 Balatonfüred paper of Erdős and Simonovits, whose display (5) proves the bound (equation_5). The same page (display (6) and the sentences before it) records the disproved conjecture that the exponent of every bipartite graph has the form 1+1/k1+1/k or 2−1/k2-1/k and the surviving conjecture that lim⁡f(n;G)/nα=c(G)\lim f(n;G)/n^\alpha=c(G) exists for some α∈(1,2)\alpha\in(1,2), "Probably the α\alpha in (6) is always rational"; these belong to Problem 713.

Source. P. Erdős, Extremal problems on graphs and hypergraphs, Hypergraph Seminar, Lecture Notes in Math. 411 (1974), 75--84; printed p. 78 = PDF p. 4 of the ten-page typescript scan (printed p. nn = PDF p. n−74n-74), read on the rendered page image. The artifact is identified in the source digest.

Read depth. Claims checked: the display and its two sentences were read clause by clause on the page image. The paper gives no proof.

Proof pointer

None in the source; see the 1970 paper's display (5).

Dependencies

None stated.

Bears on

  • Problem 576: the site's [Er74c, p. 78] source; the upper bound and the question "whether (7) is best possible", the form in which Erdős asked the cube problem in 1974.