Little topology, big volume (Q1196407)

Following Gromov say that a continuous function \(\rho: [0,\alpha) \to \mathbb{R}^ +\), \(\alpha > 0\), with \(\rho(0) = 0\) and \(\rho(\varepsilon) \geq \varepsilon\) for all \(\varepsilon \in [0,\alpha)\), is a local geometric contractibility function for a complete Riemannian manifold \(M\) if, for each \(x \in M\) and \(\varepsilon \in (0,\alpha)\), the open ball \(B(x,\varepsilon)\) is contractible in \(B(x,\rho(\varepsilon))\). The authors show that there is a function \(\nu_{\rho,n}: [0,\alpha) \to \mathbb{R}^ +\), depending only on \(\rho\) and \(n\), such that vol\((B(x,\varepsilon)) \geq \nu_{\rho,n}(\varepsilon)\) for all \(x\in M\) if \(\rho\) is a local contractibility function for \(M\) and \(\dim M = n\). This is a local version of a corresponding result of Gromov. If inj\((M) \geq 2\alpha\), then \(\rho(\varepsilon) \equiv \varepsilon\) is a function for \(M\) as above. In this case C. Croke proved vol\((B(x,\varepsilon)) \geq C_ n\varepsilon^ n\), the local version of Berger's estimate vol\((M) \geq C_ n(\text{inj}(M))^ n\).