Wiki
Wiki

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

Updated


Statement

Setting (p. 131). A 2-design on vv points is a family of lines (subsets with more than one point) such that every pair of points lies in exactly one line; see Theorem 1 for the full setting.

Theorem 2 (p. 132, quoted). "Let v=p2+p+1v=p^2+p+1. Then for p2+p+1<b<p2+2p+1p^2+p+1<b<p^2+2p+1 there is no 2-design with vv points and bb lines."

Here pp is a positive integer, not necessarily a prime or a prime power (p. 132). In the abstract's terms (p. 131), the interval [v+1,v+p−1][v+1,v+p-1] is disjoint from MvM_v.

Sharpness (p. 132). The bound is best possible whenever a projective plane of order pp exists: in such a plane replace one line {x1,…,xp+1}\{x_1,\ldots,x_{p+1}\} by the line {x2,…,xp+1}\{x_2,\ldots,x_{p+1}\} and the pp pairs {x1,xi}\{x_1,x_i\}, 2≤i≤p+12\le i\le p+1. The result is a 2-design with p2+2p+1p^2+2p+1 lines; the old line has been broken up into a near pencil on its p+1p+1 points (p. 133).

Remark (p. 132). The result fails for vv not of this form: projective planes from which points have been deleted give many examples with b−v<vb-v<\sqrt v.

Proof pointer

The paper gives two proofs. The algebraic proof (pp. 135--137) first shows that a line of more than p+1p+1 points forces a near pencil (Lemma 1, p. 135), so that all lines have at most p+1p+1 points and every point has degree at least p+1p+1 (Lemma 3, p. 136), and that a line all of whose points have degree p+1p+1 forces b=vb=v when b≤p2+2p+1b\le p^2+2p+1 (Lemma 2, p. 136). It then studies the orthogonal projection onto the complement of the row space of the v×bv\times b incidence matrix, which has rank b−vb-v, and a principal submatrix indexed by the lines through a point x0x_0 of degree p+1p+1 (which exists by Lemma 4, p. 136); if b−v≤p−1b-v\le p-1, a positivity argument forces a line through x0x_0 all of whose other points have degree p+1p+1, and Lemma 2 finishes. The combinatorial proof (pp. 138--141) derives Theorem 2 from Theorem 4, since breaking up a line of a projective plane gives at least p2+2p+1p^2+2p+1 lines by the de Bruijn--Erdős theorem.

The paper also notes (p. 133) that Theorems 2 and 3 follow from Totten's classification of the 2-designs with (b−v)2≤v(b-v)^2\le v, by a substantially longer proof.

Read depth

Claims checked: Theorem 2, the sharpness construction and the remark were read clause by clause on the page images of the print, and the algebraic proof with Lemmas 1 to 4 was followed. Nothing here is independently reviewed.

Dependencies

Theorem 4 for the combinatorial proof. External inputs named by the paper: the de Bruijn--Erdős theorem and the Stanton--Kalbfleisch bound used in Lemma 1.

Source. P. Erdős, J. C. Fowler, V. T. Sós and R. M. Wilson, On 2-designs, J. Combin. Theory Ser. A 38 (1985), no. 2, 131--142; the edition read is named on the source card.

Bears on

  • Problem 903: Theorem 2 is the problem's assertion, with blocks of at least two points: a 2-design on n=p2+p+1n=p^2+p+1 points with more than nn blocks has at least n+pn+p. The theorem needs no prime-power hypothesis on pp, and the construction above attains n+pn+p when pp is a prime power.