Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Source. Theorem 2, p. 249, of Paul Erdős and George Purdy, Some extremal problems in geometry, J. Combinatorial Theory 10 (1971), no. 3, 246--252, DOI 10.1016/0097-3165(71)90028-8, as identified on the source card.
Statement
Here is the largest number of triangles of one common positive area whose vertices are among distinct points of the plane (Section 2, p. 247; see Theorem 1 for the full notation), and is a positive absolute constant.
Theorem 2 (p. 249).
Proof pointer
Pages 249--250, sketched here. Put and take the integer points with and at most , fewer than of them. All the triangles counted have area . Given two of the points , with , the number is an integer, and if the difference of the two points is times a primitive vector, each of the lattice points on the segment between them, shifted right by , is a third vertex completing a triangle of area inside the grid (equations (3) and (4), p. 249; the paper writes for both the vector and its multiplicity). For each with the paper counts the pairs, using the density of coprime pairs in a large rectangle, and obtains more than triangles; summing over gives order , that is .
Dependencies
The asymptotic count of points with coprime coordinates in a rectangle, which the paper cites as well known (p. 250). Read depth: claims checked; the statement was read on p. 249 and the proof on pp. 249--250 for its structure only.
Bears on
- Problem 1086: a lower bound for that problem's , read as counting triangles of one common positive area. With Theorem 1 the paper leaves between and .