Abstract
We obtain a classification for the fundamental groups of closed \(n\)-manifolds of positive sectional curvature which admit an isometric \(T^k\)-action with \(k \ge \frac{n}{6}+1 (n \ne 11, 15)\).
Similar content being viewed by others
References
Bredon, G.: Introduction to Compact Transformation Groups, vol. 48. Academic Press, Dublin (1972)
Brown, K.S.: Cohomology of Groups. Springer, New York (1982)
Fang, F., Mendonca, S., Rong, X.: A connectedness principle in the geometry of positive curvature. Comm. Anal. Geom. 13(2), 479–501 (2005)
Fang, F., Rong, X.: Positively curved manifolds with maximal discrete symmetry rank. Am. J. Math 126, 227–245 (2004)
Fang, F., Rong, X.: Homeomorphic classification of positively curved manifolds with almost maximal symmetry rank. Math. Ann. 332(1), 81–101 (2005)
Fang, F., Rong, X.: Collapsed \(5\)-manifolds with pinched positive sectional curvature. Adv. Math. 221, 830–860 (2009)
Frank, P.: The Fundamental Groups of Positively Curved Manifolds with Symmetry. Thesis (2005)
Freedman, M.: Topology of four manifolds. J. Diff. Geom. 28, 357–453 (1982)
Grove, K.: Geometry of, and via symmetries. In: Univ. Lecture Ser., vol. 27, pp. 31–53. AMS, Providence (2002)
Grove, K., Searle, C.: Positively curved manifolds with maximal symmetry-rank. J. Pure Appl. Algebra 91, 137–142 (1994)
Hamilton, R.: Three-manifolds with positive Ricci curvature. J. Diff. Geom. 17, 255–306 (1982)
Hsiang, W., Kleiner, B.: On the topology of positively curved \(4\)-manifolds with symmetry. J. Diff. Geom 30, 615–621 (1989)
Kobayashi, S.: Transformation Groups in Differential Geometry. Springer, New York (1972)
Rong, X.: On the fundamental group of manifolds of positive sectional curvature. Ann. Math. 143, 397–411 (1996)
Rong, X.: Positively curved manifolds with almost maximal symmetry rank. Geom. Dedic. 59, 157–182 (2002)
Rong, X.: Fundamental group of positively curved manifolds admitting compatible local torus actions. Asian J. Math. 9(4), 545–560 (2005)
Rong, X., Wang, Y.: Fundamental group of manifolds with positive curvature and torus actions. Geom. Dedic. 113, 165–184 (2005)
Rong, X., Wang, Y.: Fundamental groups of \((4k+1)\)-manifolds with positive curvature and isometric \(T^k\)-actions are cyclic (Preprint)
Smale, S.: Generalized Poincaré conjecture in dimension \(>4\). Ann. Math. 74, 391–466 (1961)
Sugahara, K.: The isometry group of and the diameter of a Riemannian manifold with positive curvature. Math. Jpn. 27, 631–634 (1982)
Wang, Y.: On Fundamental Groups of Closed Positively Curved Manifolds with Symmetry. Thesis (2005)
Wilking, B.: Torus actions on manifolds of positive sectional curvature. Acta Math. 191, 259–297 (2003)
Wilking, B.: Group Actions on Manifolds of Positive Sectional Curvature. Universidade Federal de Goiás (2002)
Wolf, J.A.: The spaces of constant curvature. In : McGraw-Hill Series in Higher Mathematics. McGraw-Hill, Maidenheach (1976)
Author information
Authors and Affiliations
Corresponding author
Additional information
P. Frank and X. Rong are supported partially by NSF Grant DMS 1106517 and by a reach found from Capital normal university. Y. Wang is supported partially by NSFC 11001015 and 11171025.
Appendix: Proof of Theorem A for \(n<23\)
Appendix: Proof of Theorem A for \(n<23\)
For the sake of completeness and the convenience for readers, we will also present a simplified proof of the cases \(n=17, 19, 21\).
It turns out that our proof of Theorem A for \(n=17\) requires a strengthened version of (1.6.2) in [3] (see Case 3).
Theorem A1
[3] Let \(M\) be a closed \(n\)-manifold of positive sectional curvature, and let \(N_j\hookrightarrow M\) be a closed totally geodesic submanifold of dimension \(\nu _{j}(j=1, 2)\). Then the following natural homomorphisms,
are isomorphisms for \(i\le \nu _{1}+\nu _{2}-n\) and are surjections for \(i=\nu _{1}+\nu _{2}-n+1\).
Proof of Theorem A for n = 17
By Corollary 1.9, we may assume that any isotropy subgroup has fixed point set of dimension \(\le \)13.
By Lemma 1.14, we can assume two independent subgroups with torus rank at most 1, \(H_1, H_2\subset T^4\), such that \(T^4/H_i\) acts effectively on \(N_i\) up to a finite subgroup, where \(N_i\) denotes an \(H_i\)-fixed point component with \(\dim (N_i)\ge 9\). By (1.6.1), \(\pi _1(N_i)\cong \pi _1(M)\) and thus the rest of the proof is to show the desired property for \(\pi _1(N_i)\).
If \(\dim (N_1)=9\) (or \(\dim (N_2)=9)\), then \(\pi _1(N_1)\) is cyclic (Theorem 1.5). Hence, in the rest of the proof, we may assume that \(\dim (N_i)\ge 11\).
-
Case 1.
Assume that \(H_1\) is finite (or \(H_2\) is finite). If \(\dim (N_1)\le 13\) (or \(\dim (N_2)\le 13)\), then \((N_1,T^4/H)\) satisfies Theorem 1.5 and \(\pi _1(N_1)\) is cyclic. In the rest of the proof, we will assume that \(H_i\) is a circle subgroup of \(T^4\). Let \(N_0=N_1\cap N_2\).
-
Case 2.
Assume that \(\dim (N_1)=\dim (N_2)=11\). Then \(\dim (N_0)=5\) or \(7\). In the former case, \(\pi _1(N_0)\) is cyclic (Theorem 1.5), and thus \(\pi _1(N_1)\) is cyclic (1.6.1). In the latter case, we consider the isotropy representation of \(T^2\) on the normal space of \(N_0\) in \(M\). It is easy to see that there is a circle subgroup of \(T^2\) with fixed point component \(F\) of dimension \(9\) which contains \(N_0\). Then \(\pi _1(F)\) is cyclic (Corollary 1.9), and \(\pi _1(F)\cong \pi _1(M)\) (1.6.1).
-
Case 3.
Assume that \(\dim (N_1)=11\) and \(\dim (N_2)=13\). Then \(\dim (N_0)=7\) or \(9\). In the latter case, \(\pi _1(N_1)\) is cyclic (Corollary 1.9). In the former case, by Theorem A1 we may conclude that \(N_1\cap N_2\hookrightarrow N_2\) is \(7\)-connected because \(\pi _i(N_2,N_1\cap N_2)\cong \pi _i(M,N_1)=0\) for \(i\le 7\) (see (1.6.1)). Moreover, by Lemma 1.10 one concludes that the universal cover of \(N_2\) is a homology sphere, so the universal cover of \(M\) is a homology sphere (1.6.1). Then by Lemma 1.8, \(\pi _1(N_2)\) is cyclic.
-
Case 4.
Assume that \(\dim (N_1)=\dim (N_2)=13\). Then \(\dim (N_0)=9\) or \(11\). We may only consider the former case. Because \(N_0\hookrightarrow N_1\) is \(9\)-connected, \(\tilde{N}_1\) is a homology sphere (Lemma 1.10). By Lemma 1.8, \(\pi _1(N_1)\) is cyclic.
\(\square \)
Proof of Theorem A for n = 19
Without loss of generality, we may assume any isotropy subgroup has fixed point set of dimension \(\le 15\) (Corollary 1.9).
By Lemma 1.14, \(T^5\) has two independent subgroups with torus rank at most 1, \(H_1\) and \(H_2\), whose fixed point sets \(N_1\) and \(N_2\) have dimensions \(\ge 11\) on which \(T^5/H_i\)-actions are effective up to a finite group. Hence, the proof reduces to show the desired property for \(\pi _1(N_i)\) (1.6.1). Note that by Theorem 1.5 we may assume that \(\dim (N_1)=\dim (N_2)=15\). If \(H_i\) is finite, then \(5>\frac{\dim (N_1)+1}{4}\), and thus \(\pi _1(N_1)\) is cyclic (Theorem 1.5). So we may assume that \(H_i\) are circles.
If the \(T^5\)-fixed point set is not empty, then by the last part of (1.12.1), \(\pi _1(M)\) is cyclic. So we can assume that the \(T^5\)-action has no fixed point. Let \(N_0=N_1\cap N_2\). Note that \(N_0\hookrightarrow N_1\) is \(11\)-connected (1.6.2), and by Lemma 1.13 (applied to \((N_1,N_0))\), there is \(T^4\subset T^5\) with a connected fixed point set \(F\) of dimension \(3\). We then complete the proof by applying Lemma 3.2. \(\square \)
Proof of Theorem A for n = 21
Let \(H_1\) and \(H_2\) be subgroups of \(T^5\) of rank at most 1 with fixed point sets \(N_1\) and \(N_2\) of dimension \(\ge 11\) (Lemma 1.14) and such that \(T^5/H_i\) acts effectively on \(N_i\). Then \(\pi _1(M)\cong \pi _1(N_i)\) (1.6.1). Without loss of generality, we may assume that \(\dim (N_1)\le 17\) (Corollary 1.9).
If \(H_1\) is finite, then \(\pi _1(N_1)\) is cyclic (Theorem 1.5). So we may assume both \(H_1\) and \(H_2\) are circle subgroups. Since the case \(n=17\) has been proved, we may assume that \(\dim (N_i)\le 15\). If \(\dim (N_1)\le 13\), then \(\pi _1(N_1)\) is cyclic (Theorem 1.5). Hence, we may assume \(\dim (N_i)=15, i=1, 2\). Let \(N_0=N_1\cap N_2\). Clearly, we may assume \(\dim (N_0)=9, 11\) (Corollary 1.9) and \(N_0\) is fixed by a torus \(T^2\) generated by \(H_i\). If \(\dim (N_0)=9\), then \(\pi _1(N_0)\) is cyclic (Theorem 1.5) and so is \(\pi _1(N_1)\). If \(\dim (N_0)=11\), then it is easy to see, by analyzing the isotropy representation of \(T^2\) on the normal space of \(N_0\), that there is \(T^1\subset T^2\) with a fixed point set of dimension 13. By Corollary 1.9, we then conclude that \(\pi _1(N_0)\) is cyclic. \(\square \)
Note that Lemma 2.1 is not required in the above proofs.
Rights and permissions
About this article
Cite this article
Frank, P., Rong, X. & Wang, Y. Fundamental groups of positively curved manifolds with symmetry. Math. Ann. 355, 1425–1441 (2013). https://doi.org/10.1007/s00208-012-0823-z
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00208-012-0823-z