The convexity radius of a Riemannian manifold (Q525718)

A subset \(X\) of a Riemannian manifold \(M\) is strongly convex if any two points of \(X\) can be joined by a unique minimal geodesic which lies entirely in \(X\). The injectivity radius \(\text{inj}(p)\) at \(p\in M\) is defined by \[ \text{inj}(p)= \max\{R>0\,|\,{\exp_p}{|_{B(0,s)}} \text{ is injective for all } 0 < s <R\}. \] The convexity radius \(r(p)\) at \(p\) is defined by \[ r(p) =\max \{R > 0\,|\, B(p, s) \text{ is strongly convex for all } 0 < s < R \}. \] Let \(r_c(p)\) and \(r_f(p)\) be the conjugate radius and the focal radius at \(p\), respectively. Let \(\text{inj}(M) = \inf_{p\in M} \text{inj}(p)\) and similarly define the numbers \(r(M),\, r_c(M)\) and \(r_f (M)\). Given that \(M\) is compact, it is well known that the injectivity radius is characterized by the equality \(\text{inj}(M) = \min \{r_c(M), \frac{1}{2}\ell_c(M)\}\), where \(\ell_c(M)\) is the length of the shortest non-trivial closed geodesic in \(M\). The author first characterizes the convexity radius by the equality \(r(M) = \min\{r_f(M),\frac{1}{4}\ell_c(M)\}\). Then, with the aid of the last two equalities, he proves his main result: \(\inf \frac{r(M)}{\text{inj}(M)} = 0\) over the class of compact manifolds \(M\) of any fixed dimension at least two. More precisely, for each \(n\geq 2\) and \(\epsilon > 0\), there exists a compact \(n\)-dimensional Riemannian manifold M with \(\frac{r(M)}{\text{inj}(M)}< \epsilon\).