Paperlane

no. 14 · Friday 13 November 2026

Small cells in great-circle arrangements

Krishnendu Bhowmick

math.CO · 16 pp

Abstract

We study the areas of the cells determined by an arrangement $\mathcal{A}$ of $n$ distinct great circles on the unit sphere, with no three concurrent. Let $m(\mathcal{A})$ and $M(\mathcal{A})$ denote the smallest and largest cell areas, respectively. Our first result is the bound \[ m(\mathcal{A})\le \frac{C_*+O(n^{-3/4})}{n^2}, \qquad C_*=7.0666071493\ldots , \] where the implied constant is absolute. This improves the coefficient $π^2$ obtained from classical polygonal perimeter estimates. We also prove that $M(\mathcal{A})/m(\mathcal{A})>13$ for all sufficiently large $n$, improving the bound $7.43$ of L.~Fejes Tóth. The main tool is a perimeter inequality for planar polygons which takes account of their exterior angles. We combine this with a positive semidefinite kernel on the sphere and use gnomonic projection to obtain bounds for spherical cells.

comments
16 pages,3 figures
fields
Combinatorics, Metric Geometry
pdf ↗ page arXiv:2611.15244 ↗