A lower bound for the diameter of solutions to the Ricci flow with nonzero \(H^1(M^n, \mathbb R)\) (Q1429214)

Let \((M,g)\) be an \(n\)-dimensional Riemannian manifold, let \(\alpha\in H_1(M;{\mathbb Z})\), and let \(L_\alpha(g)\) be the infimum of the lengths of all curves representing \(\alpha,\) measured with respect to \(g.\) Consider the Ricci flow \[ {\partial\over{\partial t}}g=-2\text{Ric}(g),\quad 0\leq t<T. \] The authors prove the following theorem. Theorem 1. If \((M,g(t))_{0\leq t<T}\) is a compact solution of the Ricci flow, and \(\alpha\in H_1(M;\mathbb Z)\) is an element of infinite order, there exists \(c=c(\alpha,g(0))>0\) such that \(L_\alpha(g(t))\geq c\), for all \(t\in[0,T).\) By using the theorem, the authors are in a position to give an affirmative answer to a conjecture made by R. Hamilton. In fact, they prove the following corollary. Corollary 2. Let \(\overline{g}(t)=ds\otimes ds+2(n-1)(\overline{T}-t)g_{can},\) where \(g_{\text{can}}\) is a round metric on \({\mathbb S}^{n-1}.\) Then \(({\mathbb S}^1\times{\mathbb S}^{n-1},\overline {g}(t))\) cannot arise as a final time limit flow. The authors give two proofs of Theorem 1, one cohomological and one homological. In the final section of the paper, they discuss some open problems.