Wiki
Wiki

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

Updated

Erdos 1975 anti ramsey theorems

../

conjecture_1: The 1975 conjecture on the largest number of colors of an edge-coloring of the complete graph with no totally multicolored k-cycle, with the grouped coloring behind it and the authors' statement that they prove it only for triangles; the origin of the cycle question of Problem 1105.

conjecture_2: The 1975 path conjecture, its two regimes and extremal colorings, and the two theorems asserting it for n above a linear threshold and for long paths, whose proofs the paper defers and which never appeared; the origin of the path question of Problem 1105.

theorem_1: The founding anti-Ramsey limit theorem of Erdős, Simonovits and Sós: the largest number of colors on the edges of K^n with no totally multicolored copy of H, divided by n choose 2, tends to 1 - 1/d, where d + 1 is the least chromatic number of a graph obtained from H by deleting one edge.

theorem_2: The hypergraph form of the anti-Ramsey limit theorem: for a k-uniform hypergraph H, the largest number of colors on the complete k-uniform hypergraph with no totally multicolored copy of H differs by o(n^k) from the Turán number of the family of H minus one edge.

theorem_3: The structure of an extremal coloring for f(n,H): the vertices split into d classes, d as in Theorem 1, so that all but o(n^2) edges between classes carry colors used once and the edges inside each class use o(n^2) colors; the paper states it without proof.

theorem_4: The exact anti-Ramsey number of the complete graph K^p for p >= 4 and n large: one more than the Turán number of K^{p-1}, attained only by coloring the edges between d classes with distinct colors and all edges inside the classes with one further color.


P. Erdős, M. Simonovits, V. T. Sós: Anti-Ramsey theorems, Infinite and finite sets (Colloq., Keszthely, 1973; dedicated to P. Erdős on his 60th birthday), Vol. II; Colloq. Math. Soc. János Bolyai, Vol. 10, pp. 633--643, North-Holland, Amsterdam, 1975 (MR 52 #164; Zentralblatt 316.05111). The Rényi archive scan runs from printed p. 633 to p. 643, the last page being the reference list; Simonovits and Sós's 1984 reference list gives the range as 633--642.

Edition read. The copy read for this card is the Rényi archive's scan of the paper (the archive's 1975-05.pdf), eleven pages, OCR text layer from 2004; printed p. nn is PDF p. n−632n-632. The text layer garbles every formula, so the statements below were read on the rendered page images. No notice is printed in the scan (its head reads "COLLOQUIA MATHEMATICA SOCIETATIS JÁNOS BOLYAI" over "10. INFINITE AND FINITE SETS, KESZTHELY (HUNGARY), 1973." and its first and last pages carry no copyright or license line); the hosting archive's site footer speaks for the site, not the paper (https://users.renyi.hu/~p_erdos/, read 2026-10-02, prints "(C) 2005-2007 All rights reserved. All material on this site is for scientifics purposes only."); the colloquium volume has no publisher page or DOI for this edition, so the publisher's page was not consulted and no Crossref license is recorded; the term is unstated.

Read status: claims checked for the notation (printed p. 633), the statements of Theorems 1--4 with Remarks 1--2 (pp. 634--636), Lemma 1 (p. 638), Conjecture 1 with its coloring (pp. 636--637), Conjecture 2 with its extremal colorings, Remark 3, Theorems 5--6 and the closing sentence (p. 637), the second "Theorem 6" of Section 3 (p. 640), and part A of the Appendix with its short proof that f(n,C^3) = n - 1 (p. 642), read clause by clause on the page images. The proofs of Theorems 1, 2 and 4 (pp. 638--641) were read for structure only, as recorded on their pages; the proof of Theorem 2 is written out only for 3-uniform hypergraphs, and that of Theorem 4 rests on the second Theorem 6, which the paper does not prove. The proof of Theorem 3 "will not be published here", and Theorems 5 and 6 have no proof in the paper.

The paper introduces f(n,H), the largest number of colors the edges of K^n can carry with no totally multicolored (rainbow) copy of H (the site's AR(n,H)), and shows it is governed by Turán-type extremal numbers. Theorem 1 gives f(n,H)/C(n,2) -> 1 - 1/d where d+1 = min over edges e of the chromatic number of H-e; Theorem 2 is the k-uniform hypergraph analog f_k(n,H) - ext_k(n,{H-e}) = o(n^k); Theorem 3 describes the structure of extremal colorings (d vertex classes, cross edges almost all uniquely colored), with its proof withheld; Theorem 4 gives the exact value f(n,K^p) = ext(n,K^{p-1}) + 1 for p >= 4 and n large, with a unique extremal coloring on d = p-2 classes (Remark 1 relates it to Dirac's theorem). Remark 2 names the case d = 1, where Theorem 1 only gives f(n,H) = o(n^2), "degenerated", and the paper treats two such problems, cycles and paths. Conjecture 1 (p. 636) states f(n,C^k) = n((k-2)/2 + 1/(k-1)) + O(1), with the grouped coloring behind it; the authors do not assert uniqueness of the extremal colorings and say "This conjecture will be proved only for k = 3 in Theorem 5" (p. 637, as printed; Theorem 5 on that page is stated for Conjecture 2). The case k = 3, f(n,C^3) = n - 1, is proved in part A of the Appendix (p. 642), which opens "Here we prove Conjecture 2 for k = 3", a second mislabel, since its content is the triangle case of Conjecture 1. Conjecture 2 (p. 637) predicts the exact values of f(n,P^k) for k = 2t+3+ε in two ranges of n, (6) tn - C(t+1,2) + 1 + ε for n >= (5t+3+4ε)/2 and (7) C(k-2,2)+1 for k <= n <= (5t+3+4ε)/2, with the two extremal coloring types; Theorem 5 asserts it for n >= (5t+3+c)/2 with a constant c and Theorem 6 for all sufficiently large t, and the paper closes the passage with "The proofs of Theorems 5, 6, will be published later" -- both are announcements, and the site records that the proofs never appeared. A second, unrelated statement labeled Theorem 6 (an extremal number for graphs obtained from K_d(r,...,r) by adding k edges) appears in Section 3, p. 640. Method, per the paper's own account: Turán-type extremal graph theory plus a lemma relating rainbow-free colorings to the family {H-e}. This paper is the source of problem 1105: its first question is Conjecture 1 verbatim, and its second question is Conjecture 2 in the maximum form later proved by Yuan.

Source: https://users.renyi.hu/~p_erdos/1975-05.pdf.

Bears on. #1105: Conjecture 1 (printed p. 636, PDF p. 4) is the problem's cycle question and Conjecture 2 (printed p. 637, PDF p. 5) its path question; Theorems 5--6 are the results whose announced proofs the site's commentary says never appeared, and part A of the Appendix (printed p. 642, PDF p. 10) is the proof of AR(n,C_3) = n - 1 the commentary mentions. Theorem 1 (printed p. 634) gives only f(n,C^k) = o(n^2) and f(n,P^k) = o(n^2), since d = 1 for cycles and paths with k >= 3 (Remark 2, p. 636), and does not reach the linear-order values the problem asks for (page theorem_1).

Results to transcribe.

  • Theorem 1 (p. 634): With d+1 = min_{e in E(H)} k(H-e), f(n,H)/C(n,2) -> 1 - 1/d as n -> infinity (page theorem_1).
  • Theorem 2 (p. 635): For a k-uniform hypergraph H, f_k(n,H) - ext_k(n,{H-e}) = o(n^k), so the two normalized quantities share a limit (page theorem_2).
  • Theorem 3 (p. 635): In an extremal coloring the vertices split into d classes so that all but o(n^2) cross edges have their own color and each class internally uses o(n^2) colors altogether; proof not published in the paper (page theorem_3).
  • Theorem 4 (pp. 635--636): For p >= 4 and n > n_p, f(n,K^p) = ext(n,K^{p-1})
    • 1, and the extremal coloring is unique (d = p-2 classes, cross edges distinct, all inside-class edges one extra color; the hypothesis is printed "no TMC K^n", read as K^p) (page theorem_4).
  • Conjecture 1 (p. 636): f(n,C^k) = n((k-2)/2 + 1/(k-1)) + O(1), via a coloring using n/(k-1) groups of k-1 vertices; "proved only for k = 3" as printed, the proof being part A of the Appendix (p. 642) (page conjecture_1).
  • Conjecture 2 / Theorems 5--6 (p. 637): Exact values of f(n,P^k) for k = 2t+3+ε with the two extremal coloring types; asserted for n >= (5t+3+c)/2 (Theorem 5) and for all large t (Theorem 6), with proofs that "will be published later" (page conjecture_2).

No file of this source is held: no license on record permits its redistribution, and the card cites the edition it names above.