Ball growth at all scales in universal covers of short metric graphs
Abstract
Let Γ be a connected metric graph of first Betti number b≥ 2, and let Γb be a trivalent graph of the same Betti number with all edges of unit length. We show that if length(Γ) ≤ λ length(Γb), then the universal cover of Γ contains a vertex around which, \emph{at every radius} R≥ 0, the ball is at least as long as the ball of radius R in the trivalent tree whose edges all have length 3λ. In particular, for λ≤1/3, this ball is at least as long as the ball of the same radius in the universal cover of Γb, with no multiplicative loss. This improves an earlier theorem of the author, which gives the weaker lower bound (1−3λ) times the ball in the universal cover of Γb, and only for λ≤1/6. The proof rests on an elementary comparison principle. Among trees whose vertices have degree at least three and whose edges have length at most c, the trivalent tree with all edges of length c has the smallest balls at every radius. As an application, we double the exponential growth rate in the author's earlier uniform lower bound for the area of balls in the universal cover of small-area surfaces of hyperbolic type, from ln 2 to 2ln 2, beyond the hyperbolic rate. We also obtain explicit constants in the positive answer to Guth's question for surfaces. Finally, we conjecture the sharp, all-scale version of the minimal volume entropy inequality of Kapovich--Nagnibeda and Lim, and show that it would be optimal.
Related
- Finite Cyclic Fundamental Groups and Exotic Universal Covers from the Cartwright--Steger Surface
- Exact fillings using universal covers in Floer theory
- Improved bounds for universal convex covers of unit arcs
- On $4$-perfect-matching covers of cubic graphs with two adjacent odd circuits in a $2$-factor
- A new infinite family of covers of complete graphs that admit a half-transitive action of a Suzuki simple group
BibTeX
@misc{karam2026ball,
title = {{Ball growth at all scales in universal covers of short metric graphs}},
author = {Steve Karam},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.20088},
eprint = {2610.20088},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.20088},
}
RIS
TY - UNPB
TI - Ball growth at all scales in universal covers of short metric graphs
AU - Steve Karam
PY - 2026
DA - 2026/10/14
PB - arXiv
DO - 10.48550/arxiv.2610.20088
UR - https://arxiv.org/abs/2610.20088
AB - Let Γ be a connected metric graph of first Betti number b≥ 2, and let Γ_b be a trivalent graph of the same Betti number with all edges of unit length. We show that if length(Γ) ≤ λ length(Γ_b), then the universal cover of Γ contains a vertex around which, \emph{at every radius} R≥ 0, the ball is at least as long as the ball of radius R in the trivalent tree whose edges all have length 3λ. In particular, for λ≤1/3, this ball is at least as long as the ball of the same radius in the universal cover of Γ_b, with no multiplicative loss. This improves an earlier theorem of the author, which gives the weaker lower bound (1−3λ) times the ball in the universal cover of Γ_b, and only for λ≤1/6. The proof rests on an elementary comparison principle. Among trees whose vertices have degree at least three and whose edges have length at most c, the trivalent tree with all edges of length c has the smallest balls at every radius. As an application, we double the exponential growth rate in the author's earlier uniform lower bound for the area of balls in the universal cover of small-area surfaces of hyperbolic type, from ln 2 to 2ln 2, beyond the hyperbolic rate. We also obtain explicit constants in the positive answer to Guth's question for surfaces. Finally, we conjecture the sharp, all-scale version of the minimal volume entropy inequality of Kapovich--Nagnibeda and Lim, and show that it would be optimal.
ER -