Extremal points of infinite clusters in stationary percolation (Q1962217): Difference between revisions

In a stationary percolation on the complete set of edges in \(Z^d\) the author defines a ``very-special'' cluster \(C\) by the following conditions: 1) In some given coordinate direction -- which we fix as the first coordinate, for convenience, \(C\) has infinitely many lowest points; that is, points \(z\) such that \(z(1)=\min(z'(1): z' \in C)\). 2) There is an infinite path in the opposite direction in \(C\); that is, \((z_1,z_2,\dots)\) such that \(\lim_{n\to\infty} z_n(1)=\infty\). He proves that there are almost surely no very-special clusters in a stationary two-dimensional percolation with bounded range (the range being defined as the length of the longest possible open edge). He also gives counterexamples to show that both conditions, dimension 2 and finite range, are required. In addition, he shows that even under the stated conditions merely ``special'' clusters may exist, that is, clusters with infinitely many lowest points, but which are merely unbounded (without an infinite path in the first-coordinate direction).