Splitting off tori and the evaluation subgroup of the fundamental group (Q1825467)

The author studies the following problem: Given a space X, what is the largest torus \(T^ n\) such that X is homotopy equivalent to \(Y\times T^ n?\) In this paper all spaces are connected CW complexes. For such a space X its toral number is n if n is the largest integer such that X is homotopy equivalent to \(Y\times T^ n\) for some space Y. The author remarks that Y is necessarily homotopically equivalent to a covering space of X. Let G be a subgroup of \(\pi_ 1(X)\) and h(G) the image of G under the Hurewicz homomorphism h in the homology group. Then h(G) may contain free summands of \(H_ 1(X)\), and the Hurewicz rank of G is the maximum rank of these free summands. If there is no free summand in h(G), then the author says that the rank of G is zero, and if there is no maximum, then the Hurewicz rank of G is infinite. The main result of the paper is the following: Theorem. The toral number of X is equal to the Hurewicz rank of \(G_ 1(X)\) (Gottlieb's evaluation subgroup of the fundamental group). Corollary 1. There exists a finite covering \(\tilde X\) of X whose toral number is greater than or equal to the rank of \(h(G_ 1(X)).\) Corollary 2. The dimension of X is greater than or equal to the rank of \(G_ 1(X).\) Corollary 3. If F is dominated by a finite complex and is the fibre of a fibration with simply connected total space, then F is homotopically equivalent to a finite complex except possibly when F has a nontrivial finite fundamental group. Corollary 4. (Conner-Raymond Splitting Theorem [\textit{P. E. Conner} and \textit{F. Raymond}, Topology 10, 283-296 (1971; Zbl 0236.57023)]). If a torus \(T^ k\) acts on X so that evaluation at a point gives a map \(\omega\) : \(T^ k\to X\) so that \(\omega_*(H_ 1(T^ k))\) is a direct summand of \(H_ 1(X)\) of rank k, then X is equivalently homeomorphic to \(T^ k\times Y\), for some space Y, where \(T^ k\) acts on the product by \(g(h,y)=(gh,y).\) As the author specifies, this paper was inspired by a question of Haefliger [\textit{A. Haefliger} and \textit{E. Salem}, Riemannian foliations on simply connected manifolds and actions of tori on orbifolds, Preprint (1988)], to which Corollary 3 is a simplified answer and it owes its existence a key observation of Siebenmann. The main result of the author is related to a result of \textit{W. Lück} [J. Pure Appl. Algebra 45, 143- 169 (1987; Zbl 0657.57010)]. The author affirms that he has the information that \textit{J. Oprea} [A homotopical Conner-Raymond theorem and a question of Gottlieb, Can. Math. Bull. (to appear; Zbl 0661.57014)] has independently discovered a version of the main theorem of this paper and its relation to the Conner-Raymond splitting theorem.