We derive high-density percolation for uniform canonical hard-disk packings from Georgii's equivalence of ensembles and Magazinov's high-activity percolation theorem. For every epsilon > 0, every weak limit of uniform square-torus packings sufficiently close to the maximum number density 1/(2 sqrt(3)) has an infinite component when centers at distance at most 2+epsilon are joined. The parameter conversion uses the explicit bound rho <= 1/(2 sqrt(3)) - exp(-4 pi lambda)/9 for stationary Gibbs packings of activity lambda. We give the ensemble argument for arbitrary square sizes and treat the maximum-density endpoint separately. The conclusion addresses the canonical percolation question of Bowen, Lyons, Radin and Winkler; it is not a crystallization theorem. Separated triangular patches show why the uniform sampling hypothesis cannot be replaced by high density alone. This AI-assisted research note is self-audited and unrefereed, credits its published theorem inputs, and makes no absolute priority claim.
No takes yet. Share an insight, caveat, or question.
Alper Ferudun (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: