Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Statement
Setting (p. 248, Section 1). For points of the plane, is the minimum, over all planar sets of points, of the number of different distances the set determines. The paper notes (the equilateral triangle), and .
Theorem 1 (p. 248, quoted). "The minimum number of distances determined by points of a plane satisfies the inequalities"
Here is a constant the paper does not specify.
Higher dimensions (p. 248, the paragraph closing Section 1). For points in -dimensional space, with the corresponding minimum, the paper states that "the same method yields" . No proof is written out and the constants are not specified.
The paper says (p. 248) that it has sought to improve Theorem 1 for many years without success.
Source. P. Erdős, On sets of distances of points, Amer. Math. Monthly 53 (1946), 248--250; Theorem 1 and its proof on p. 248. The copy read is identified on the source card.
Read depth. Claims checked: the statement, the setting and the -dimensional remark were read clause by clause on the page images, and the proof was followed. Nothing here is independently reviewed.
Proof pointer
P. 248. Lower bound. Take a vertex of the convex hull of the points. If the distances from to the other points take values and the most frequent of them occurs times, then . The points at that distance lie on one semicircle about , so their distances to the first of them are distinct values. Hence , which is least when ; this gives . Upper bound. The integer points with number at least , and their distances are of the form with ; the number of integers up to that are sums of two squares is less than , which the paper takes from Landau's Verteilung der Primzahlen, vol. 2.
Dependencies
Landau's count of the integers up to that are sums of two squares (the paper's footnote, p. 248).
Bears on
- Problem 89: the problem asks whether every points in the plane determine distinct distances. The theorem's upper bound, from the integer grid, shows that this order could not be improved; its lower bound is .
- Problem 1083: the -dimensional remark gives , stated without proof; the problem asks, for , whether the minimum is , the exponent of the remark's upper bound.